Introduction

Transcription

Introduction

Investigations into ring shake
of chestnut
Patrick Fonti
DISSERTATION
N° 14732
WSL Swiss Federal Research institute
Sottostazione Sud delle Alpi
Bellinzona
ETH Swiss Federal Institute
of Technology
Zürich
Bellinzona-Zürich 2002
I
Diss. ETH No. 14732
Investigations into ring shake of chestnut
A dissertation submitted to the
SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZÜRICH
for the degree of
DOCTOR OF NATURAL SCIENCES
presented by
Patrick Fonti
Dipl. Forst. Ing. ETH
Swiss Federal Institute of Technology Zürich
born 14 September 1971
from Miglieglia TI, Switzerland
accepted on the recommendation of
Prof. Dr. Jürgen Sell, examiner
Prof. Bernard Thibaut, coexaminer
Dr. Nicola Macchioni, coexaminer
2002
III
The thesis-report (Section A) is based on the results presented in the
following six papers, referred to in the text by their Roman numerals, and
which are annexed in Section B. Discussion of specific results regarding
each single study are enclosed in the respective papers.
I.
Fonti, P., Macchioni, N., and Thibaut, B. (2002): “Ring shake in
chestnut (Castanea sativa Mill.): state of the art.” Annals of forest
sciences, Vol. 59 (2), 129-140.
II.
Fonti, P., and Macchioni, N. (submitted 2002). “Ring shake in
chestnut: anatomical description, extent and frequency of failures”.
Annals of forest sciences.
III.
Fonti, P., Bräker O-U., and Giudici, F. (2002): “Relationship
between ring shake incidence and earlywood vessel characteristics in
chestnut wood”. IAWA Journal, Vol. 23 (3), 287-298.
IV.
Fonti, P., and Frey, B. (2002). “Is the ray volume a possible factor
influencing ring shake occurrence in chestnut wood?”. Trees –
Structure and function, Vol. 16 (8), 519-522.
V.
Fonti, P., and Sell, J. (in press). “Radial split resistance in chestnut
earlywood and its relation to the ring width”. Wood and fiber
science.
VI.
Fonti, P. (submitted 2002). “Growth strain and ring shake in chestnut
coppice trees”. Forêt méditerranéenne. Translated in French.
Accepted papers are published with the kind permission of the journals
concerned.
V
Table of contents
Section A
1
SUMMARY
3
RIASSUNTO
4
INTRODUCTION
5
Background
New potentialities for chestnut timber
The problem of ring shake
Past research activities
General framework
5
6
7
7
8
OBJECTIVES
9
STUDY DESIGN
10
MATERIALS AND METHODS
12
SYNTHESIS OF CHIEF RESULTS
14
Origin and typologies of chestnut ring shake
Evidence of a peculiar radial mechanical weakness of chestnut wood
Indications of a mechanical self-optimisation of trees
Growth stresses as the principal “driving force” responsible for ring shake
Influence of growth on radial wood strength
Control of ring shake risk
14
16
17
18
20
20
CONCLUSIONS
21
ACKNOWLEDGEMENTS
22
REFERENCES
22
Section B
25
1. RING SHAKE IN CHESTNUT (CASTANEA SATIVA MILL.): STATE OF
THE ART
27
2. RING SHAKE IN CHESTNUT: ANATOMICAL DESCRIPTION, EXTENT
AND FREQUENCY OF FAILURES
45
3. RELATIONSHIP BETWEEN RING SHAKE INCIDENCE AND
EARLYWOOD VESSEL CHARACTERISTICS IN CHESTNUT WOOD
59
4. IS THE RAY VOLUME A POSSIBLE FACTOR INFLUENCING RING
SHAKE OCCURRENCE IN CHESTNUT WOOD?
73
5. RADIAL SPLIT RESISTANCE IN CHESTNUT EARLYWOOD AND ITS
RELATION TO THE RING WIDTH
83
6. GROWTH STRAIN AND RING SHAKE IN CHESTNUT COPPICE TREES
95
Section A
1
Section A
Section A
3
SUMMARY
This work is concerned with the development of ring shake, a wood defect that very
often occurs in European chestnut (Castanea sativa Mill.). This study must be placed in
a general research context aimed at assembling a better picture of the complex
phenomenon of ring shake in order to evaluate new preventive measures that will
minimise the risk of occurrence. The main objectives are (1) the examination of the
relationships between radial wood cohesion, growth stresses and ring shake occurrence,
and (2) an analysis of the effect that improved growth due to new silvicultural concepts
aimed at quality coppice wood production has on ring shake occurrence.
Investigative studies were carried out on wood material collected in chestnut coppice
forests from the southern part of the Swiss Alps. Analyses were performed on the basis
of a comparison of wood with and without ring shake.
Anatomical investigations of the ring shake privileged plane of fracture (earlywood
area) showed that ring shaken wood displays a higher amount of rays and smaller
earlywood vessel lumina. Mechanical rupture testing revealed that earlywood radial
split resistance is positively correlated with ring width. Examination of longitudinal
strain-induced displacement performed on the stem surface of coppice trees did not
reveal the existence of a direct relationship between high stress level and ring shake
occurrence. Lastly, through a review of literature combined with a description and
quantification of the different typologies of chestnut ring shake so far observed,
important elements for the interpretation of results were collected.
It has been highlighted that the principal cause of ring shake in chestnut is associated
with the unbalanced interrelation between the weak structure of the tangential wood
plane and the release of growth stresses. Stress overloaded trees react by producing a
strengthened wood. When trees are cut however, mechanically originated ring shakes
develop as a consequence of the relieving of growth stresses. Nevertheless mechanical
testing proved that not all trees are prone to ring shake in the same way. This result is
particularly encouraging because, compatible with the new silvicultural methods, it is
then possible for the forester to operate in a manner which reduces the risk of ring
shake.
4
Section A
RIASSUNTO
Questa ricerca indaga sul problema della cipollatura, un difetto del legno che spesso
insorge nel castagno europeo (Castanea sativa Mill.). Il presente studio si inserisce in
un contesto più ampio di ricerca volto alla comprensione del complesso fenomeno della
cipollatura. L’intento finale è quello di sviluppare nuove misure preventive che
permettano di ridurre il rischio di cipollatura. Gli obiettivi principali della tesi sono (1)
lo studio delle relazioni fra coesione del legno, tensioni di crescita e incidenza della
cipollatura, e (2) la valutazione dell'effetto che una crescita radiale più regolare e
sostenuta, conseguente all’introduzione di nuovi tecniche selvicolturali mirate a una
produzione di legname di qualità, ha sullo sviluppo della cipollatura.
Gli studi investigativi sono stati effettuati su materiale legnoso raccolto in boschi cedui
castanili situati nella regione Svizzera a Sud delle Alpi. Le analisi sono state eseguite in
base ad un confronto tra legno con e senza cipollatura.
Le indagini anatomiche, concentratesi nella zona in cui la cipollatura si sviluppa
maggiormente (legno primaverile), hanno indicato che il legno cipollato si caratterizza
da una maggiore quantità di raggi e da vasi del legno primaverile più piccoli rispetto al
legname sano. Le prove di rottura meccanica hanno rivelato che la resistenza alla
fenditura radiale del legno primaverile è correlata positivamente con la larghezza
dell'anello. L'esame delle tensioni sulla superficie del fusto di alberi cresciuti in boschi
cedui non ha consentito di verificare direttamente l'esistenza di un rapporto fra un alto
livello di tensioni e la presenza di cipollatura. Infine, grazie a uno studio bibliografico
sul tema e ad una descrizione e quantificazione delle diverse tipologie di cipollatura
finora osservate nel castagno, sono stati raccolti importanti elementi per
l'interpretazione dei risultati. Si è quindi potuto evidenziare che nel castagno la causa
principale della cipollatura è riconducibile al disequilibrio tra la debole struttura
tangenziale-longitudinale del legno e il rilascio delle tensioni di crescita. Gli alberi
sovraccarichi di tensioni reagiscono producendo un legno più tenace. Quando tuttavia
questi vengono tagliati, il rilascio delle tensioni di crescita causa comunque cipollature
di origine meccanica. Ciononostante, i test di rottura meccanica hanno mostrato che non
tutti gli individui sono inclini allo stesso modo alla cipollatura. Questo risultato è
particolarmente incoraggiante in quanto, compatibilmente alle nuove tecniche
selvicolturali, è possibile operare a favore di un rischio ridotto di cipollatura.
Section A
5
INTRODUCTION
Background
European chestnut (Castanea sativa Mill.) is the only native species of the genus in
Europe. Its centre of origin is distributed in several areas of the Southern Mediterranean
Basin. Humans have long known how to manage the chestnut in extremely profitable
and diversified ways. The artificial diffusion of the chestnut in the past went far beyond
the natural range of the species, becoming capillary diffused mainly in rural and
mountainous areas, where it constituted an abundant agro-forestry resource (Pitte, 1986;
Bernetti, 1987). Even at the present time chestnut is one of the more commonly
occurring deciduous tree species of the Mediterranean Basin and of the central
European area. Throughout Europe chestnut is present in at least 25 countries and
covers a total area of 2.5 million ha. Concerning the Switzerland territory, chestnut is
mainly present in the southern part of the Swiss Alps, where it constitutes 21% in
number of the all trees (Brändli and Brassel, 1999).
The importance that the chestnut has gained for centuries is principally connected to its
use as a flexible, multi-purpose species: this species can be managed as coppice forests,
as highforest or as a grove for the production of a large set of products, ranging from the
traditional wood (durable timber for both high quality and bio-engineering use, biomass
for energy, tannins, etc.) and non wood products (fruits, litter, grazing, honey, etc).
Since the Roman age and about until the second World War chestnut coppice forests,
due to the high stool resprouting ability, the lasting vitality of the root system and the
high productivity, have been primarily managed for timber production. However the
traditional management system applied, based on short rotations (12-20 years) and
thinnings, was at first directed at the production of a large variety of row assortments
(principally poles) used externally because of the high durability of the wood.
Subsequently, since the late 1940s, the progressive decline of the rural economy and the
onset of the fungus Cryphonectria parasitica, a pathogen of bark cancer, caused an
increasing indifference for the cultivation of chestnut forests (Pitte, 1986), which led to
the reduction of maintenance practices. In abandoned chestnut coppices the stands tend
to become overaged, invaded by other species and to lose their functionality and
productivity (Amorini et al. 1997; Conedera et al. 2001). The timber production has
been reduced in both quantitative and qualitative terms.
6
Section A
In recent times, as a result of the increasing consciousness of a sustainable behaviour
and the release of pathologies, the social and environmental functions of the chestnut
resource have assumed a political relevance in the development policies of marginal
rural areas: landscape improvement, biodiversity protection, maintenance of cultural
tradition, tourism development, fire prevention, hydro-geological regulation, soil
protection etc. are all compatible with chestnut management practices.
New potentialities for chestnut timber
This renewed interest is providing new perspectives for chestnut cultivation. With
particular reference to these and specifically to the production of timber, this species has
four undeniable advantages that place it in a very favourable position: it is (i) capillarydiffused in rural and mountainous areas, (ii) a fast-growing tree-species which displays
(iii) a valuable and appreciated timber that (iv) can be produced in conformity with the
principles of sustainability (Bourgeois, 1992). The high natural durability, the good
technological characteristics and the appreciated aesthetic properties of this wood make
it suitable for a variegate range of natural products ranging from more innovative
(parquet floor, veneer, cabinet work, joinery) to more traditional ones (poles, fences,
vineyard stakes, energy).
However traditional management systems are no longer fully adequate in responding to
new market requirements demanding larger sized stems of good quality in order to
allow for easier and more rational processing.
First results based on modern silvicultural methods compatible with the principles of
sustainable development seem however to prove (from tests carried out so far in welldefined and specific conditions) that it is technically possible and economically feasible
to meet this market quality requirement. The basic principles are the extension of
rotation times up to 30 or 50 years and the application of early and frequent thinnings of
moderate-heavy intensity (Bourgeois, 1992, Amorini et al. 2000; Manetti et al. 2001).
The principal goal of the treatments in this case is the production of high quality wood
assortments because of their economical “driving force”. However, other differentiated
products, for example poles, can also be complimentarily produced over the whole
productive cycle.
Section A
7
The problem of ring shake
The renewed process of valorisation and exploitation of this natural and renewable
resource is however running up against some difficulties, which obstruct its
implementation. Among these, there is in particular the extreme tendency of chestnut
wood to develop ring shake: a type of wood fracture arising parallel to the annual
growth rings in the tangential plane of the trunk (Chanson et al., 1989), which strongly
reduces the portion of useable wood and consequently its economical value.
Unfortunately, due to the high risk of ring shake occurrence and to the consequent low
production yield and more laborious processing, the interest in adopting silvicultural
alternatives for improving the production of quality timber remains slight. Minimising
the risk of ring shake would represent a fundamental step towards making the
management of chestnut coppices economically attractive. This would stimulate, where
production circumstances are adequate, a self-regulated management process. In this
way a relevant part of coppice areas adapted for good quality and high value timber
production could be restored and at the same time the development of rural areas would
be promoted.
Past research activities
Research on chestnut ring shake has been predominantly directed towards
understanding the phenomena that engender ring shake in order to find or develop
preventive or remedial measures capable of reducing the risk and effects of occurrence.
The first research activities on the topic began in the 1980s, particularly in France
(Ferrand, 1980; Thibaut, 1982). Initial studies were mainly descriptive in character. As a
result of these investigative efforts it was possible to identify some important aspects of
the problem, in particular the distinction of two basically different types of ring shake:
the "healthy" and the "traumatic" (Chanson et al., 1989). While the causes of traumatic
ring shake are quite clear – simply scar tissue laid upon dead tissue to which it does not
bond – the phenomenon of "healthy" ring shake is much more complex and therefore
also more difficult to approach.
Subsequently many different studies have attempted to define, quantify and describe the
problem (Ferrand, 1980; Bonenfant, 1985; Cielo, 1988; Boetto, 1991; Macchioni, 1992;
Macchioni and Pividori, 1996, Pozzi, 1996). In recent years further studies have focused
particularly on the identification of the factors involved in the development of ring
8
Section A
shake. The goal of these studies was the estimation of the risk of ring shake (Frascaria et
al., 1992; Elzière, 1995). In spite of the limited approaches adopted (analysis of single
interrelations in univariate models) some hypotheses on the possible indirect causes
have already been formulated. These include genetic factors (Frascaria et al., 1992),
regularity of growth (Fioravanti, 1992; Amorini et al., 1998), silvicultural treatment
(Elzière, 1995) and the technological characteristics of chestnut wood (Leban, 1985;
Chanson, 1988; Frascaria et al., 1992; Macchioni 1995). These studies were later
tentatively compiled in FOREST, a co-ordinated pan-European programme proposing
first preventive and remedial measures. Nevertheless, since ring shake is a highly
complex phenomenon, a distinct identification of the causes leading to ring shake as
well as guaranteed measures against its occurrence are not yet available.
General framework
From a merely mechanical point of view, failures always result from an unbalance
between strength and stresses. In the case of the development of ring shake there are in
particular wood stresses acting in the radial direction that exceed the wood tangentiallongitudinal plane strength-capacity (I). The phenomenon of ring shake can therefore be
considered as the sole and last consequence resulting from the intricate action of
numerous factors that affect the equilibrium between radial wood strength and stress.
Because of the complexity of defining the role played by each single factor in the ring
shake process, it has been decided to primarily focus the investigations on these two
pivotal elements, i.e. characterising the strength-stress relationship. In this way we hope
to collect useful knowledge permitting us to identify which circumstances are necessary
for the onset of ring shakes. It will then be easier to revisit the factors at the origin of
such a situation. The major difficulties in this study lie however in the quantification of
both of these pivotal components that change in space and time (as the tree grows in
size or when wood is processed), establishing a dynamic and complex relationship
between them. In fact, on the one hand wood strength can vary greatly within and
between annual rings, stem portions and individual trees, and is principally associated
with wood moisture. On the other hand there are stresses acting in wood originating
from various causes and which can appear at different times and operate in a combined
way.
Section A
9
In order to come across this difficulties, studies attempting to characterise strength and
stress of chestnut wood will be accompanied by studies on chestnut wood anatomy.
Such an approach should indirectly supply much important information on the
susceptibility of the wood specimen to ring shake because of the evident relationship
existing between anatomy and transverse tensile strength (Beery et al. 1983;
Schniewind, 1959). These investigative studies were principally based on a comparison
approach among chestnut individuals with and without ring shake grown under the same
stand conditions and which have been manipulated in the same way. The collection of
homogeneously selected material permits us to disregard the exposure condition (which
should be the same) whereas the approach based on antagonism should allow an easier
identification of the differences among shaken and unshaken individuals.
OBJECTIVES
This research must be placed in a more ample context aimed at improving the
production and exploitation of quality chestnut wood from coppice forests. This thesis
should be seen as an additional contribution towards understanding the phenomena that
engender healthy (mechanically originated) ring shake in chestnut. In this case
investigations mainly deal with the hypothesis of a structural wood weakness in relation
to the internal stresses, which are considered to be the principal causes of the extremely
high ring shake susceptibility of chestnut.
The principal objectives pursued are:
•
the collection of new basic knowledge into the relationship between radial wood
cohesion, growth stresses and ring shake occurrence;
•
the verification of the effect that growth factors, following the application of modern
silvicultural methods, have on ring shake occurrence.
In particular this research project intends:
1. to collect specific information relative to the occurrence of healthy ring shake;
a) collecting, organising and summarising already available knowledge on the
phenomenon of chestnut ring shake;
b) describing the different fracture morphology, time-occurrence and frequency of
the different ring shake typologies;
10
Section A
2. to verify whether among the chestnut individuals there are differences in
characteristics relevant for the radial wood strength-stress equilibrium that could
explain why some individuals develop healthy ring shake while some others, grown
under the same stand conditions, do not;
a) assessing whether ring shaken wood displays an anatomically different wood
structure which might favour an increased wood radial weakness;
b) developing a new split testing procedure to obtain comparable split resistance
measurements;
c) verifying whether ring shaken wood is effectively weaker than the unshaken
one as well as the effect that the ring width has on wood radial strength;
d) verifying whether ring shaken trees have effectively higher growth stresses
than the unshaken ones;
3. to assess whether faster radial growth favours the development of ring shake;
4. to propose silvicultural measures in order to reduce the risk of ring shake.
STUDY DESIGN
The project was divided into three major parts (Figure 1). In the introduction a short
review of the general aspect of chestnut ring shake is given. In particular this part
consisted of a collection of fragmentary and sometimes difficult to acquire knowledge
in order to obtain, as much as possible, a complete, correct and up-dated picture of the
phenomenon of chestnut ring shake. All this knowledge has been collected, organised,
discussed and summarised in a state of the art paper (I).
The main, or investigative, part of the thesis was in turn subdivided into single tasks that
include independent investigation blocks. A first precious step towards understanding
the phenomenon of ring shake has been made due to a descriptive study on typologies,
occurring-time and frequency of chestnut ring shake (II). Further studies were
essentially based on the investigation of the existence of relationships between some
selected anatomical, mechanical and rheological characteristics of chestnut wood and
the incidence of ring shake. These investigation blocks were focused on:
•
the quantification of wood anatomical characteristics of earlywood vessels (III) and
radial rays (IV) that are supposed to influence the radial strength properties of
wood;
Section A
•
11
the characterisation of the mechanical radial strength properties of wood derived
from measurements of the earlywood radial split resistance (V);
•
the characterisation of longitudinal strain induced deformation measured on the
stem periphery of standing chestnut individuals (VI).
The conclusive part, which is presented in the next sections of the present report
(synthesis of chief results), attempts to interpret, evaluate and summarise the various
aspects studied throughout this dissertation project. In addition, based on the newly
acquired knowledge, measures aimed at reducing the risk of ring shake occurrence are
Introduction
proposed.
Section B
I
Ring shake in chestnut
(Castanea sativa Mill.):
State of the Art
II
Investigations
Ring shake in chestnut:
anatomical description, extent
and frequency of failures
III
IV
Relationship between ring shake
incidence and earlywood vessel
characteristics in chestnut wood
V
VI
Radial split resistance of
chestnut earlywood and its
relation to the ring width
Conclusions
Is ray volume a possible factor
influencing ring shake occurrence
in chestnut wood?
Growth strain and ring shake in
chestnut coppice trees
Section A
Synthesis of chief results
Figure 1. Schematic representation of the study design.
12
Section A
MATERIALS AND METHODS
The wood material used for the investigative studies was collected in chestnut coppices
from the Southern part of the Swiss Alps. The choice of the plots was influenced by the
availability of areas with single coppicing. Six plots and a total of 110 trees were
considered (Bedano 20, Bedigliora 24, Brione 21, Gerra 20, Gorduno 5 and Novaggio
20). Since the production of quality wood is the goal, analysed individuals were selected
from among the overstory trees. As described in Fonti et al. (1998) ring shake
characterisation was performed on wood discs collected from the base of tree stems,
where ring shake occurs with major frequency (I). Anatomical, mechanical and
rheological characteristics were surveyed as close as possible to the ring shake survey.
Such measurements are however somewhat specific to the purposes of this work and
therefore the methods had to be partly adapted to our specific needs.
Thanks to recent advances in image analysis techniques comparative studies on
quantitative anatomical features are today relatively easy to perform. Nevertheless,
study materials should be prepared adequately in order to obtain sufficient quality
images allowing a correct identification and quantification of the features being
measured. While for earlywood vessels the measure was directly possible on images
captured from the previously polished wood surface (III), the procedure for rays was
more laborious because of the superior magnification required as well as for the
difficulties in their identification. For scanning an electron microscope was used and
rays were manually transferred from images to a transparent before automatic
computational analysis (IV).
The characterisation of the radial strength of chestnut wood has already been attempted
in several studies (Macchioni 1995; Frascaria et al. 1992; Leban 1985). Nevertheless,
these earlier studies suffered from some limitations in the testing methods, because they
do not allow either comparable data to be collected or the radial cohesion strength of
wood to be properly characterized. Tensile radial tests performed by Leban (1985) only
allowed one measurement per radial specimen, without it being known exactly where
the specimen would break. Bending tests carried out by Frascaria et al. (1992) permitted
several measurements along the same radial specimen, but again in this case it was not
possible to check the exact crack location in this case. Macchioni (1995) developed a
testing method based on torsion load on wood cores. This method permitted multiple
Section A
13
measurements along the same cores and enabled the load location to be checked.
However, the fracture displayed features differing from ring shake, not occurring
precisely in the tangential plane. The wedge splitting technique patented by Tschegg
(1986) permits the radial split resistance to be determined while avoiding previous
limitations. In particular, as opposed to previous attempts, the wedge splitting technique
allows us, due to a starter-notch, to previously determine the exact location, even within
the earlywood zones, where the fracture will occur. However, the shape and dimensions
of the test specimens are unsuitable for easy measurement of the split resistance of
annual rings positioned in close proximity. After several attempts, a split testing method
adequate for our requirements was achieved (V). This consists of a simplified Tschegg’s
(1986) wedge splitting technique that is easier to perform and adapted to such a
comparative study.
The quantification of stresses is instead somewhat complicated, because they are not
directly measurable but only calculable from the measurable strain components and
modulus of elasticity. In an experimental approach strain is often used as a parameter
for growth stresses, because it can be specified relatively precisely. The basis of this
measurement is the following: a material will be abbreviated, when it is compressed and
elongated, if tensile forces are applied. If the material is released from tensile or
compression stress, it tries to recapture the original dimension. The difference between
the dimension of stressed and non stressed material is called strain. As described by
Archer (1986) stress can be measured on the wood surface or within the xylem. A
simple method of strain measurement called “single hole drilling method”, developed
by Fournier et al. (1994) is particularly attractive. In this case the structure of the fibres
are interrupted by a hole drilled on the outer part of the stem to get a local longitudinal
relaxation of the wood, which is measured by a dial gauge. Compared to the xylem
measurement, this measurement has the principle advantage that it can be practised at
standing trees with relatively low destructive effects. So it is possible to quantify the
original stress level under a condition of genesis while the tree is still standing and ring
shake has not already had the opportunity to develop, which could bias the survey. The
study on the quantification of stresses (VI) was therefore based on the indirect
measurement of longitudinal strain measured conformably by the methods proposed by
Fournier et al. (1994).
14
Section A
It is also important to highlight that, although the methods used may not be ideally
suited to accurately identify the characteristics pursued, nevertheless given that the
investigation was based on the contrast of opposing situations and that the
measurements were always performed with the same procedure, the information gained
should however be adequate for our investigative requests.
Figure 2 supplies a short overview of objectives, materials and methods of each single
study carried out within this work.
Part
Study
Specific objectives
•
Introduction
I
Investigations
II
Investigations
III
Investigations
IV
Investigations
V
collect, organise and summarise
already available knowledge on the
phenomenon of chestnut ring shake
•
describe the different fracture
morphology, time-occurrence and
frequency of the different ring shake
typologies
• assess whether ring shaken wood
displays an anatomically different
structured wood which might favour
an increased wood radial weakness
• assess whether ring shaken wood
displays an anatomically different
structured wood which might favour
an increased wood radial weakness
• develop a new split testing
procedure to obtain comparable split
resistance measurements
• verify whether ring shaken wood is
effectively weaker than the
unshaken one as well as the effect
that the ring width has on wood
radial strength
•
Investigations
VI
Conclusions
Section
A
•
Methods
Analysis (collection,
organisation, evaluation and
interpretation) of already
available knowledge
Journal articles, Conference proceedings,
Reports and Thesis
Descriptive observation of ring
shake
45 wood discs from 2 mature coppice
stands (Brione and Bedigliora).
Survey of quantitative earlywood
anatomical characteristics and
analysis of the relationship to
ring shake occurrence
Survey of quantitative radial rays
anatomical characteristics and
analysis of the relationship to
stands (Novaggio, Gerra, Bedano). 10
shake-free and 10 with extreme ring shake
occurrence from each selected stand.
stands (Novaggio, Gerra, Bedano). 10
shake-free and 10 with extreme ring shake
occurrence from each selected stand.
Pith-to-bark radial strips (at least 2 for
each selected trees) from 50 wood disc
gathered in 3 differently managed
coppices stands (Gorduno (5 trees):
regularly and intensively thinned; Brione
(21): occasionally and partially managed;
Bedigliora (24): unmanaged and
abandoned)
50 selected trees from 3 differently
managed coppices stands (Gorduno (5
trees): regularly and intensively thinned;
Brione (21): occasionally and partially
managed; Bedigliora (24): unmanaged and
abandoned)
Survey of earlywood radial split
resistance and analysis of the
relationship to ring width and
Survey of longitudinal stressverify whether ring shaken trees have induced displacement on stem
surface and analysis of the
effectively higher growth stresses
than the unshaken ones
relationship to ring shake
occurrence
propose, as far as possible,
preventive measures in order to
reduce the occurrence of ring shake
in chestnut
Material
Analysis (organisation,
interpretation and evaluation) of
collected information
Studies I to VI
Figure 2. Overview of objectives, materials and methods.
SYNTHESIS OF CHIEF RESULTS
Origin and typologies of chestnut ring shake
By definition healthy (mechanically originated) ring shakes, in opposition to the
traumatic ones, are failures that appear to be unrelated to any recognisable anatomical
perturbation (Chanson et al., 1989). Looking more closely at ring shaken failure
Section A
15
surfaces it has been observed that, at an anatomical level, separations show different
features depending on their origin. In chestnut three types of ring shake have been
observed: the “overlay”, “detachment” and “crack” typologies (I, II). Each of these
typologies is the specific result of an established unbalance between strength and
stresses. The first one, which has a clear traumatic origin, is characterised by scar tissue
that is merely superposed on dead cells without any connection between them. In this
case the wood has no strength capacity. The origin of the detachment-type is not clear
yet. The separation mainly develops at a ring boundary as a debonding in the middle
lamella layer and does not exhibit direct association with damages, even if an indirect
effect could not be excluded. The strength of the middle lamella layer is likely reduced
and if enough stress is applied a debonding develops. The last type observed is the crack
typology, a separation that arises across the cell walls of the earlywood vessels (Figure
3). This type, which is not related to specific annual rings nor to a traumatic event, has a
clear mechanical origin (II). In this case however, whether a failure opens is due either
to a structural weakness and/or to an extremely high stress value.
Given that this last type is the most frequent failure type in chestnut wood, which
accounts for 84% of the total ring shaken failures surveyed (II) and is therefore one of
the foremost causes of wood quality losses, the present research has mainly
concentrated on this failure type.
Figure 3. Enlarged cross section of chestnut wood affected by a crackform ring shake.
16
Section A
Evidence of a peculiar radial mechanical weakness of chestnut wood
Chestnut wood tends to develop tangential (i.e. concentric) splits while most of the
other wood species form radial fractures. This particular behaviour of chestnut wood
and the fact that it is the species most commonly affected by ring shake suggest that in
the radial direction this type of timber might be particularly weak compared to other
species. In fact the wood structure of chestnut is formed in such a way that makes it,
compared to other wood species, specifically susceptible to tangential failure (I). The
earlywood zone, which has plenty of large vessels, represents a naturally weak zone and
can consequently easily fail. Some species, as beech, cope with this weakness due to
numerous and large wood rays, which through a high radial tensile strength are able to
reinforce the earlywood zone (Schniewind, 1959; Beery et al., 1983; Mattheck et al.,
1994; Badel and Perré, 1999; Burgert et al., 1999; Burgert and Eckstein, 2001). But this
is only partially the case for chestnut wood, which is characterised by small and thin
uniseriate wood rays and thus cannot offer great resistance to failures developing in the
tangential plane (Figure 4).
Figure 4. Structure of chestnut wood. Characteristic of chestnut are
the ring porous wood structure with large earlywood vessels and the
thin monoseriate radial rays.
Section A
17
Clearer evidence of the existence of a weak tangential plane in chestnut wood structure
is revealed by its high ring shake frequency, in particular of the crack-form (II). The
wood structure along the tangential plane may be too weak to resist mechanical stresses,
therefore explaining the high susceptibility of chestnut to ring shake. Unfortunately
earlywood radial strength data comparing chestnut wood with other wood species are
lacking, thus making it difficult to confirm this belief. Nevertheless, comparing radial
split resistance among chestnut trees with and without ring shake showed that the defect
occurs more frequently in weak growth rings (V).
Indications of a mechanical self-optimisation of trees
Analysis of wood structural anatomical characteristics between chestnut wood with and
without ring shake have shown that, contrary to expectations, ring shaken wood displays
a greater radial rays volume (IV) and smaller earlywood vessels lumina (III) than the
unshaken ones. Hence we reject the hypothesis that among chestnut trees ring shake is
favoured by the weakening effect of larger earlywood cell lumina and by a reduced
amount of radial rays.
This surprising result could however be founded if biomechanical processes stressing
the tree are taken into consideration. In particular, during their lifecycle trees are
subjected to several mechanical stresses (growth, gravitationally and environmentally
induced stresses) that affect the growth. Consequently the tree forms its wood in order
to carry out both its biological and physical functions in an optimal way (Mattheck and
Kubler, 1995). In this case it is plausible that, in order to confront the problem of ring
shake, stress-overloaded trees might generally create more rays and smaller earlywood
vessels lumina than the stress-poor ones. In other words, to avoid undesired fractures,
the wood must be provided with bearings that compensate for these stresses through an
enhanced strength. As described by Mattheck and Kubler (1995) and proved by
Albrecht et al. (1995) and Dietrich (1995) the strength of wood depends on the
distribution of the mechanical stresses that act in the tree. Through processing these
stresses can be relieved. Stress-overloaded trees have therefore a higher risk of
developing wood failures. The expected outcome is consistent with those observed in
the present studies (III, IV), even if apparently contradictory.
18
Section A
Growth stresses as the principal “driving force” responsible for ring
shake
Growth stresses develop in each stem and in each branch of a tree. These are the result
of the superposition of support stresses and maturation stresses. Support stresses are
caused by the self-weight supported by the tree. Maturation stresses originate on
growing wood cells, which tend to contract in the fibre direction, and to expand
transversely against restraining forces of adjoining older wood cells (Kubler, 1987).
Tree felling and crosscutting of stems into logs releases longitudinal growth stresses
near the cut, so that logs change dimension at the new ends (Figure 5).
Periphery
Crosscut
Pith
+
+
-
-
Figure 5. Example of stress relieving due to crosscutting. Above:
exaggerated dimensional changes in a log as a consequence of the
release of longitudinal growth stresses. The core expands while the
periphery contracts. This cross-cut strain triggers ring shake in
chestnut wood. Below: distribution of longitudinal growth strain in
stems. Tension areas marked +; compression areas -.
Section A
19
The redistribution of internal stresses caused by the relief of residual stresses in cut
surfaces is capable of producing stress concentration effects, and is the likely cause of
splits and shakes observed in freshly cut timber (Kubler, 1987). Subsequently,
throughout manufacturing, the wood can be subjected to drying or other processes that
engender additional stresses that contribute to the development of supplementary
failures.
Observations performed on the timing ring shake develops have shown that green wood
ring shakes represent 70% of the entire observed shakes, while the remaining 30%
developed during the drying process particularly concentrated in those wood discs that
were already affected by the defect in the green state (II). The timing fractures occur
(II), which is generally prior to timber processing, and the position they often have (I)
let us presume that the set off mechanism leading to ring shake must be related to
stresses acting on the standing tree or on the fresh felled stems.
The observations performed in the anatomical studies (III, IV) indicate that growth
stresses are associated with the development of ring shake. In fact stress-overloaded
trees create a more resistant wood. When however the trees are felled and the internal
wood balance is broken, some of the stress can be released (Archer, 1986; Kubler,
1987) and new cracks can develop. The stress-overloaded trees then have more stress to
release, which in some cases is so strong that the enlarged volume of rays and the
reduced earlywood vessels lumina may not be able to counteract. The results are
therefore as observed, i.e. ring shaken wood is characterised by smaller earlywood
vessels and by a larger amount of rays (III, IV).
As mentioned already, these observations highlight therefore former hypotheses that the
mechanism leading to the crack separation is principally related to the relieving of wood
growth stresses (Chanson, 1982; Thibaut et al., 1995).
The study relating growth strain-induced displacement on standing trees and ring shake
occurrence on stem wood discs did not reveal any clear relationship between ring shake
occurrence and the high longitudinal surface strain (VI). However, due to some
limitation of the applied method in estimating stress levels, it is not possible to assess or
reject whether ring shake incidence is related to elevated growth stresses
20
Section A
Influence of growth on radial wood strength
Measurements of earlywood radial split resistance performed on chestnut wood have
shown that radial split resistance and annual ring width are positively correlated and that
rates of ring shake occurrence increase in narrow and weak growth rings (V). This,
however, does not mean that all narrow rings or slow-growing trees are always affected
by ring shake. In fact, besides radial strength, also growth stresses in wood are very
important with respect to the development of ring shake, even if a direct relationship
between ring shake and growth stresses could not be directly assessed (VI). It can,
therefore, likely happen that ring shake occurs in wood displaying a high radial strength
and vice versa. However, an increase of radial growth corresponds to an improvement
of the radial split resistance, which plays an important role in the development of ring
shake. Therefore, an enhancement of the radial wood cohesion should help to reduce the
risk of ring shake occurrence.
Another aspect that has to be considered is the regularity of growth. It has been
observed that the occurrence of ring shake is not only concentrated in a limited number
of rings which are characterised by a current narrow radial increment followed by larger
rings, but also in trees that have grown irregularly (Amorini et al., 1998; Fonti, 1997;
Macchioni and Pividori, 1996; Chanson, 1988). These changes in growth rhythm create
heterogeneity in the wood material where stresses can concentrate engendering in these
zones a supplementary risk of ring shake.
Control of ring shake risk
Considering the aforementioned aspects, and in order to limit the ensuing damage of
ring shake, the following could be taken into account wherever quality chestnut wood
(with a lower risk of ring shake) is to be produced. While on one hand foresters can
influence the radial wood cohesion by enhancing and guaranteeing a fast-growth (V), on
the other hand it is important to limit growth stresses as much as possible. An
equilibrated tree shape is therefore preferred, but Kubler (1987) recommends avoiding
stimuli for reorientation because of the risk of causing new supplementary stresses.
All of this can be obtained in accordance with the new proposed silvicultural concepts
through both an adapted selection of site (favouring fertile soils) and an active
silviculture. Through early and frequent thinning competition decreases and trees can
Section A
21
better profit from the resulting free-spaces. As a consequence trees will grow more
regularly, more quickly attaining the minimum dimension requested by the market. In
doing so radial strength will be improved, reorientation stresses diminished and the
period of exposure to risk reduced.
While these measures do not completely exclude the possibility of ring shake
occurrence, they certainly minimise both the risk and the consequences of ring shake.
CONCLUSIONS
This research provided further insight in understanding the phenomenon of chestnut
ring shake. Newly acquired knowledge on the stress-strength relationships supports and
completes findings obtained in former studies conducted on the same topic. In particular
it was highlighted that in European chestnut the basic mechanism triggering ring shake
primarily lies in the imbalanced relationship between the radial wood cohesion and the
release of growth stresses. Whereas by the overlay and the detachment typologies
failures open due to the reduced strength capacity of wood, by the crack typology,
which is largely the most frequent in chestnut, failures are associated with a structural
weakness peculiar to the wood structure. In fact the earlywood vessels generate a
privileged environment for ring shake fractures that the thin and uniseriate radial rays
are not able to counteract.
Several studies on beech provide indications that trees facing higher stresses during
their lifecycle produce a wood that is optimised and adapted to deal with these
importunities. Results obtained from comparative analyses of anatomical characteristics
of chestnut wood with and without ring shake let us suppose that trees react by
producing wood that displays more rays and smaller earlywood vessel lumina. When
trees are felled or crosscut into logs, blocked stresses near the cut will release. Usually
the relieving of stresses becomes manifest through radial checks or twists. Chestnut
wood structure however favours the opening of tangential failures rather than radial
ones. Stress overloaded trees denote, immediately after felling, the development of a
major incidence of ring shake.
Nevertheless, not all chestnut trees have the same tendency to be ring shaken. Studies
on the relationship between growth factors and ring shake occurrence have in particular
shown that reduced radial growth corresponds to weaker wood which in turn tends to
display an increased amount of ring shake.
22
Section A
An enhancement of the radial wood cohesion, which is compatible with new
silvicultural concepts, should therefore help in reducing the risk of ring shake
occurrence.
ACKNOWLEDGEMENTS
The present study has been carried out thanks to the financial support of the “WSL
Swiss Federal Research Institute”, the “Swiss Forest Agency”, the “Forest Services of
both canton Ticino and Grigioni”, the “Dipartimento Istruzione e Cultura”, the
“Federlegno Ticino” and the “Fondazione Legno, Lions Club Bellinzona e Moesa”. I
am especially thankful to Marco Conedera and Fulvio Giudici who accompanied me
throughout this PhD with precious advice and support; to Prof. Ladislav Kucera (†) and
Jürgen Sell who were willing to lead the research project; to Bernard Thibaut and
Nicola Macchioni for their precious advice and collaboration; to Franco Fibbioli for
field assistance; to Christine Favre for text-revision and to all other colleagues who
made contributions during field and institute work.
REFERENCES
Albrecht, W. A., Bethge, K. A., and Mattheck, C. (1995). “Is lateral strength in trees controlled
by lateral mechanical stress?” Journal of Arboriculture, 21(2), 83-87.
Amorini, E., Bruschini, S., Fioravanti, M., Macchioni, N., Manetti, M. C., Thibaut, B., and
Uzielli, L. (1998). “Studi sulle cause d'insorgenza della cipollatura nel legno di castagno
(Castanea sativa Mill.).” Convegno nazionale sul castagno, Cison di Valmarino
(Treviso), 269-292.
Amorini, E., Bruschini, S., and Manetti, M. C. (1997). “Conversion into high forest of chestnut
coppices: hypothesis for an alternative and sustainable management.” XI World Forestry
Congress, Antalaya, Turkey 13-22 October 1997.
Amorini, E., Chatziphilippidis, G., Ciancio, O., Di Castri, F., Giudici, F., Leonardi, S., Manetti,
M. C., Nocentini, S., Pividori, M., Rapp, M., Romane, F., Sevrin, E., and Zingg, A.
(2000). “Sustainability of chestnut forest ecosystems: Is it possible?” Ecologia
mediterranea, 26(1-2), 3-14.
Archer, R. (1986). Growth stresses and strain in trees, Springer Verlag Berlin, 240.
Badel, E., and Perré, F. E. (1999). “Détermination des propriétés élastiques d'éléments
individuels du plan ligneux du chêne par des essais de traction sur micro-éprouvettes.”
Ann Sci For, 56, 467-478.
Beery, H., Geza, I., and Mc Lain, E. T. (1983). “Quantitative wood anatomy - relating anatomy
to transverse tensile strength.” Wood and Fiber Science, 15(4), 395-407.
Bernetti, G. (1987). I boschi della Toscana, Giunta regionale Toscana, Edagricole, Bologna,
177.
Section A
23
Boetto, G. (1991). “La cipollatura del castagno: osservazioni in due boschi cedui del Piemonte
(Chiusa Pesio CN, Bibiana TO),” Tesi di laurea, Univesità degli studi di Torino, Torino,
104.
Bonenfant, M. (1985). “Croissance et qualité du châtaignier de futaie en Bretagne”, 123.
Bourgeois, C. (1992). Le châtaignier: un arbre, un bois. IDF, Paris, 376.
Brändli, U.B., and Brassel, P. (Red) (1999). "Commento ai risultati. In:Inventario forestale
Nazionale Svizzero: Risultati del secondo inventario 1993-1995. Birmensdorf, Istituto
federale di ricerca per la foresta, la neve e il paesaggio. Berna, Ufficio federale
dell'ambiente, delle foreste e del paesaggio. Berna, Stoccarda, Vienna, Haupt, 27-36.
Burgert, I., Bernasconi, A., and Eckstein, D. (1999). “Evidence for the strength function of rays
in living trees.” Holz als Roh- und Werkstoff, 57, 397-399.
Burgert, I., and Eckstein, D. (2001). “The tensile strength of isolated wood rays of beech (Fagus
sylvatica L.) and its significance for the biomechanics of living trees.” Trees, 15, 168170.
Chanson, B. (1982). “Approche structurale et histophisiologique du bois de châtaignier: Etude
prélimianire a celle de la roulure.” DEA en sciences foresrtières, 170.
Chanson, B. (1988). “Etude de la variabilité de quelques propriétés physique et anatomique du
bois de châtaignier (Castanea sativa Mill). Application à l'étude de la roulure,” Thèse,
Université des sciences et techniques du Languedoc Montpellier, Montpellier, 278p.
Chanson, B., Leban, J.-M., and Thibaut, B. (1989). “La Roulure du châtaignier.” Forêt
méditerranéenne, 11(1), 15-32.
Cielo, P. (1988). “Incidenza e tipologie della cipollatura in un ceduo di castagno del comune di
Garessio (CN): ricerche bibliografiche e sperimentali,” tesi di laurea, Università di
Torino, Torino, 356.
Conedera, M., Stanga, P., Oester, B., and Bachmann, P. (2001). “Different post-culture
dynamics in abandoned chestnut orchards and coppices.” For. Snow Landsc. Res., 76(3),
in press.
Dietrich, F. (1995). “Wie der grüne Baum tangentiale Zugbelastungen bewältigt,” Dissertation,
Universität Karlsruhe, Karlsruhe, 119.
Elzière, S. (1995). “Influence de l'eclaircie sur des brins de taillis de châtaignier (Castanea
sativa Mill.): Analyse de tige, contraintes de croissances et fissilité du bois,” tesi di
laurea, Ecole nationale d'ingenieurs des travaux agricoles de Bordeaux, Bordeaux, 57.
Ferrand, J.-C. (1980). “La roulure du châtaignier (Castanea sativa Mill.). Note préliminaire.”
Seichamps (France CNRF), 2, 17.
Fioravanti, M. (1992). “Caratterizzazione del legno giovanile di castagno (Castanea sativa
Mill.): studi su anatomia, densitometria e variazioni dimensionali,” Dottorato, Università
degli studi di Firenze, Firenze, 75.
Fonti, P. (1997). “Studio delle correlazioni tra alcune caratteristiche del soprassuolo e dei
polloni di castagno (Castanea sativa Mill.) con l'incidenza e la distribuzione della
cipollatura in un ceduo a Novaggio, Ticino,” Tesi di laurea, ETHZ, Zurigo, 88p.
Fonti, P., Giudici, F., Kucera, L. J., Ott, E., and Pöhler, E. (1998). “Studio sulla cipollatura in un
ceduo castanile.” Convegno nazionale sul castagno, Cison di Valmarino (Treviso), 23-25
ottobre 1997, 293-302.
Fournier, M., Chanson, B., Thibaut, B., and Guitard, D. (1994). “Mesures de déformation
résiduelles de croissance à la surface des arbres en relation avec leur morphologie.
Observations sur différentes espèces.” Annals of Forest Science, 51, 249-266.
Frascaria, N., Chanson, B., Thibaut, B., and Lefranc, M. (1992). “Génotypes et résistance
mécanique radiale du bois de châtaignier (Castanea sativa Mill.). Analyse d'un des
facteurs explicatifs de la roulure.” Ann Sci For, 49, 49-62.
24
Section A
Kubler, H. (1987). “Growth stresses in trees and related wood properties.” Forest Products
Abstracts, 10(3), 61-119.
Macchioni, N. (1992). “Studi sulla cipollatura del castagno (Castanea sativa Mill.): metodologie
per la valutazione della coesione trasversale del legno,” Dottorato, Università degli studi
di Firenze, Firenze, 102.
Macchioni, N. (1995). “Mechanical strength and ring shake in chestnut (Castanea sativa Mill.).”
Forêt Méditerrannéenne, 16(1), 67-73.
Macchioni, N., and Pividori, M. (1996). “Ring shake and structural characteristics of a chestnut
(Castanea sativa Miller) coppice stand in northern Piedmont (northwest Italy).” Ann Sci
For, 53, 31-50.
Manetti, M-C., Amorini, E., Becagli, C., Conedera, M., and Giudici, F. (2001). “Productive
potential of chestnut (Castanea sativa Mill.) stands in Europe.” For. Snow Landsc. Res.,
76(3), in press.
Mattheck, C., Albrecht, W., and Dietrich, F. (1994). “Die Biomechanik der Holzstrahlen.”
Allgemeine Forst- und JagdZeitung, 165, 143-147.
Mattheck, C., and Kubler, H. (1995). Wood - The internal optimization of trees, Springer Verlag
Berlin Heidelberg, 129.
Leban, J.-M. (1985). “Contribution à l'étude de la roulure de châtaignier,” Thèse de doctorat,
I.N.P.L., Nancy, 164p.
Pitte, J. R. (1986). Terres de castanide. Hommes et paysages du châtaignier de l'Antiquité à nos
jours. Librarie Arthème Fayard, 479.
Schniewind, A. P. (1959). “Transverse anisotropy of wood: a function of gross anatomic
structure.” Forest products journal, 9, 350-359.
Thibaut, B. (1982). “Mise en place d'un programme de recherches sur la roulure du
châtaignier.” Bulletin de vulgarisation forestière (IDF), 12-15.
Thibaut, B., Fournier, M., and Jullien, D. (1995). “Contraintes de croissance, recouvrance
différée à l'étuvage et fissuration des grumes: cas du châtaignier.” Forêt
Méditerranéenne, 16(1), 85-91.
Section B
25
Section B
Section B
27
1
Ring shake in chestnut (Castanea sativa Mill.):
State of the art
Patrick Fontia, Nicola Macchionib, Bernard Thibautc
a
WSL Swiss Federal Research Institute, Sottostazione Sud delle Alpi,
Via Belsoggiorno 22, Casella postale 57, CH-6504 Bellinzona. Switzerland
b
Istituto per la ricerca sul legno, Consiglio nazionale delle ricerche,
Via A. Barazzuoli 23, I-50136 Firenze, Italy
c
Université Montpellier II Science & Techniques du Languedoc,
Laboratoire de Mécanique & Génie Civil, Place Eugène Bataillon,
Bât. 13 - F-34095 Montpellier, France
Annals of Forest Science (2002) 59(2): 129-140
Submitted: 22 January 2001
Accepted: 1 October 2001
Published: March 2002
28
Section B
Abstract
Too often chestnut wood (Castanea sativa Mill.) becomes economically uninteresting
because of the high risk of ring shake to which this species is prone. For more than
twenty years chestnut ring shake has been the subject of studies undertaken in an effort
to understand its underlying causes and mechanisms. Since not all aspects of the
phenomenon have been sufficiently studied at the present time, ring shake has not yet
been completely elucidated. However, it is possible to outline a general framework of
the phenomenon and advance preliminary ideas on the causes that contribute to the
development of this type of fracture. This article summarises the current state of
knowledge, discusses the possible causes and proposes measures to reduce the risk of
ring shake occurrence in chestnut.
Ring shake / Castanea sativa Mill. / wood / residual stresses / mechanical strength
Résumé
La roulure du châtaignier (Castanea sativa Mill.): connaissances actuelles. Trop
souvent le bois de châtaignier (Castanea sativa Mill.) perd son intérêt économique à
cause du haut risque de roulure qui affecte cette espèce. Depuis plus de vingt ans la
roulure du châtaignier fait l’objet de plusieurs études vouées à la compréhension des
causes et des mécanismes qui conduisent à sa formation. A l'état actuel tous les aspect
n’ont pas étés suffisamment étudiés pour que l’on puisse considérer la roulure comme
un phénomène complètement élucidé. Malgré cela, un cadre général du phénomène peut
être esquissé et des premières réflexions sur les causes qui mènent à l’apparition de ce
type de fracture peuvent être avancées. Cet article résume l'état des connaissances
acquises à ce jour, discute les causes possibles et propose des mesures afin de diminuer
le risque d'apparition de la roulure chez le châtaignier.
Roulure / Castanea sativa Mill. / bois / contraintes résiduelles / résistance
mécanique
Section B
29
Introduction
Chestnut (Castanea sativa Mill.) is widespread in about 15 Mediterranean and Central
European countries with a total cover of more than 2 million hectares [12]. Until the
mid-20th century, chestnut was of fundamental importance to the economy and to the
subsistence of rural populations. Then, with the decline of the rural economy and the
onset of diseases, the management of chestnut forests ceased. However, chestnut timber
possesses a pleasant appearance, high durability and good mechanical properties. Since
it can be processed using modern manufacturing or industrial techniques (laminated,
veneer, lumber, non-structural Glulam and solid wood panels) suitable for such addedvalue sectors as furniture, equipment and carpentry, it is one of the most versatile and
appreciated woods growing in Europe [12]. One of the main problems to be taken into
account is the risk of ring shake, whose occurrence greatly reduces the value of the
timber assortment. In the worst case, the incidence of ring shake is so high that only few
logs of a stand can be brought to the sawmill. With ring shake as the main obstacle to
the economical exploitation of chestnut wood, today's forest managers are not ready to
invest in chestnut forests. As a result chestnut wood tends to be a largely neglected
natural renewable resource.
Ring shake is a widespread phenomenon affecting a great number of species of both
softwood and hardwood and is found in trees grown in temperate and tropical climates.
In general, however, it afflicts only a very small proportion of trees. Irrespective of
whether the cracks occur after felling or cross cutting, they are nearly always radial
cracks. Some species are more heavily prone to ring shake occurrence (such as some
species of the genus Quercus, Juglans, Abies, Pseudotsuga, Tsuga and Eucalyptus
[19]), but chestnut is probably the most widely affected species, since it is nearly
impossible to find a forest plot without any ring shaken log.
Research into ring shake is aimed at understanding the factors that cause the fracture in
order to evaluate new preventive measures that will minimise the risk of occurrence.
This would permit the reintroduction of a "driving force" for chestnut forest
management.
This review summarises the fragmentary acquired knowledge currently available about
ring shake in chestnut wood and discusses the causes of the phenomenon. We also
propose measures for decreasing the risk of ring shake occurrence. In doing so we
intend to open up new discussions of the subject and provide a solid base of knowledge
for future investigations.
Fundamentals of chestnut ring shake
Definition
At the end of the 1980s Chanson [16] and Cielo [19] published objective definitions of
ring shake, distinguishing between a description of the phenomenon and its causes.
Discarding indications of the causes or the processes that lead to ring shake, since none
of the several suggested explanations was widely accepted, they simply defined ring
shake by its appearance, i.e. a separation in the tangential plane that occurs in the
ligneous tissues along the annual growth ring.
Annals of forest science (2002), Vol. 59 (2), 129-140
30
Section B
Where and when ring shake appears
Ring shake occurs mainly in stem wood. In some cases it can also appear in the big
branches of aged trees, but this is quite rare. It usually does not occur in roots. Opinions
diverge on whether ring shake is already present in standing trees, with several authors
believing that it might be at least partially present in living trees [19, 22]. Radial
ultrasonic measurement of stems evidenced that waves propagate more slowly in stems
which displayed ring shake immediately after felling [35]. This may be due to a break in
wave propagation caused by the fracture. But other factors, such as a decrease in the
radial moduli of elasticity of trees prone to ring shake, could also explain the slowing of
wave propagation [51]. Conversely, other authors [18] believe ring shake to be found
immediately after felling results from the releasing of growth stresses in the stem when
the stem is crosscut. At all events, new ring shake may occur after the cutting of the tree
either as a result of the logs being dried and sawn or even as a result of the installation
of wood products [18]. After felling, two different trends of crack development and
propagation are observed on the logs, depending on the occurrence of ring shake
displayed by the freshly felled stem. If the stem displays some ring shake after felling,
the wood drying and wood heating process tends to increase its size or number;
conversely, if the stem displays radial cracks rather than ring shake, logs cut from it are
inclined to form and extend radial cracks [1, 10, 18, 19, 25]. In some cases newly
formed radial cracks may also change direction, going off a tangent to generate new
ring shake [51].
Types and features of fractures
In his observations Chanson [17] distinguished two types of ring shake: "traumatic" ring
shake and the more common "healthy" ring shake. By definition, "traumatic" ring shake
is always related to visible anomalies in the wood tissue, whereas with "healthy" ring
shake, splitting appears to be unrelated to any recognisable anatomical perturbation.
Two fracture features can be distinguished in traumatic ring shakes (Figure 1): the first,
which we have called "overlay", is characterised by scar tissue superposed on dead cells
without any connection between them [16]. It is also possible for ring shake to arise
indirectly as a consequence of trauma. In this case a process of compartmentalisation
appears, leading to discoloration and decay in the surrounding area [22], and ring shake
may occur in the wood tissue as a detachment between the anomalous cells. We called
this second feature "discoloured detachment". The "healthy" type also displays other
fracture features, depending on the manner in which the wood cells are separated
(Figure 1). A first feature of observed ruptures is detachment in the compound middle
lamella layer between cells. This kind of shake mainly develops at a ring boundary [61]
and is typical of ring shake caused by wood drying [58]. We named this feature
"detachment". The second feature of fractures that occur predominately in ring shake
developed in fresh green wood immediately after the tree-felling consists of a crack that
arises across the cell walls of the earlywood vessels [58]. This latter feature was called
"crack" by virtue of evidence of a break in the wood cell wall opposing it to the
detachment feature, where the material seems more to be “unglued”.
Ring shake in chestnut: state of the art
Section B
31
Overlaya
Discoloured detachmentb
Image
Features
Type
Traumatic
Detachmentc
Crack d
Image
Features
Type
Healthy
Figure 1. Types and features of fractures viewed in cross section. a Overlay of new cells on dead tissues due to a
physiological reaction to wound on cambium after trauma. Appears on standing trees. b Discoloured detachment
between cells at the ring boundary between the earlywood zone of the annual ring and the latewood zone of the
previous ring; develops as a consequence of compartmentalisation after trauma leading to discoloration and to local
decay with detachment. May develop in standing trees. c Detachment between cells at the ring boundary between the
earlywood zone of the annual ring and the latewood of the previous one. Mainly characteristic of ring shake that
appears as the wood dry. d Crack across cell walls in the earlywood zone. Appears predominately in fresh green
wood.
32
Section B
Distribution and incidence
In relationship to environmental and anthropogenic factors
Several authors have investigated the relationship between environmental factors and
ring shake. One of these is Chang [15], who, in his general review of ring shake in
different species, pointed out that ring shake is not determined by a unique element, but
it is rather the result of several factors that act together. One of these factors might be
temperature, since it is hypothesised that frost or sudden temperature change might open
fractures in wood. Observations on chestnut stands partly support this hypothesis [41].
According to Chang, chestnuts growing in cold zones seem to be more affected by ring
shake than those growing in temperate zones. This theory is contested by results
obtained by Boetto [10] and Cielo [19], which showed that exposition and elevation
have no effect on ring shake intensity.
A further element thought to play a role in the ring shake process is soil. In his study on
ring shake in oak, Lachaussée [48] reported that the problem occurs less frequently in
trees growing in fertile soils than in trees growing in poor ones. Observations reported
in other studies on chestnut support this hypothesis, even though the differences in the
incidence of ring shake are only slight [4, 41, 57].
In addition to environmental conditions, anthropogenic factors have also been reported
to be involved in the ring shake process. In fact, although the defect is present in every
management system, be it coppice stand, high forest (plantation or natural) or orchards,
it was observed that the risk of ring shake in chestnuts growing in high forests is minor
[18, 22, 41]. In addition, a recent study by Amorini et al. [4] revealed that in coppice
stands a positive relationship exists between a high silviculture intensity and a lower
risk of ring shake formation.
In the chestnut distribution area
As yet there has been no comprehensive study of ring shake propagation across all areas
in which chestnut is grown, but many indicators of occurrence in mature stands allow us
to deduce that in the Mediterranean area at least, ring shake occurs wherever chestnut
grows. Investigations in different areas of southern France give indications of regional
differences for both “traumatic” (more frequent in Mediterranean regions) and “healthy”
ring shake (more abundant in Limousin then in Perigord, for example) [57].
Fragmentary investigations conducted principally in France and Italy support this belief
(Table I), even if ring shake is quite rare in some localised areas.
Among chestnut trees
Several authors have undertaken analyses in an effort to identify tree characteristics that
will enable us to differentiate ring shaken trees from unshaken ones. In general, it
proved very difficult to identify such characteristics for trees grown on the same stands:
in practice ring shake not only occurs both in trees that display an equilibrated
morphological structure and in trees that do not [10, 17, 19], but also in dominant and
dominated trees [4, 10, 19, 54]. Likewise, bark morphology and chestnut blight
(Cryphonectria parasitica) do not seem have any impact on fracture development [19].
However, a few authors observed that old and/or big trees might be more inclined to
develop ring shake [1, 10, 17, 19, 22, 41, 54]. Results from further investigations using
the multivariate analysis method bear out this trend. [17, 25, 54].
Section B
33
Table I. Incidence of ring shake observed in various studies
Stand location
Sample
Surveying Method
LanguedocRoussillon (F)
94 shoots taken
from 9 stands
Observations on two increment
cores
Brétagne (F)
480 shoots taken
from 24 stands
Observation at the base of the
logs
Languedoc,
Roussillon Limousin
and Perigord (F)
Pyrénées Orientales,
Cévennes, Limousin
and Périgord (F)
Aude, Aveyron,
Hérault, Lot, Tarn,
Tarn and Garonne (F)
Piemonte (I)
Piemonte (I)
% ring shaken stem per plot/region
on green wood a
after drying b
60
(from 17 to 90 depending on
stand)
40
(from 5 to 100 depending on
stand)
Source
[51]
-
[11]
[57]
285 shoots taken
from 37 stands
logs
42
-
[18]
156 shoots taken
from 6 regions
logs
39
-
[18]
45 shoots taken
from 3 stands (15
shoots each)
82 shoots taken
from 3 regions
Observation on 5 cm thick disks
taken at different heights, starting
from the base of the logs
taken from the bases of 300
shoots
taken from the bases of 93 shoots
53
53
[19]
-
36
[1]
-
38.5
[10]
-
38 (of shoots)
96 (of standard)
[54]
-
40
[4]
54
68
[24]
Piemonte (I)
50 shoots taken
from 2 regions
Piemonte (I)
0.3 ha coppice
stand
Toscana, Lazio and
Piemonte (I)
35 shoots taken
from 4 stands
Ticino (CH)
0.1 ha coppice
stand
a
Ring shake observed immediately or a few days after the felling of the tree
b
Ring shake observed on dried wood (<15%)
Analysis of ring shake occurrence within a single coppice stand indicates that the
phenomenon is not randomly distributed throughout the tree population, but is instead
concentrated over a number of stools. In particular it was observed that the incidence of
ring shake among all shoots of the same stool tends to be the same [25, 54]. Considering
the ring shake incidence of the standards (shoots that stay for two rotation periods), this
appears somewhat remarkable. In fact Macchioni and Pividori [54] observed in their
study that all the standards displayed ring shake, even if all the other shoots in the same
stool did not. The authors also observed that ring shake in standards mainly occurs near
the annual rings, corresponding to the years of the cut of the previous coppice stand.
Within trunks
In general, it has been noticed that longitudinal ring shake occurs mainly at the base of
the stem [1, 4, 10, 17-19, 51], while radial ring shake (from the pith to the bark) is
distributed with unimodal frequency in the middle third of the radius [1, 10, 11, 25, 54].
It has been observed that drying increases ring shake intensity and that the new
distribution is slightly shifted towards the bark [25]. The defect appears to be randomly
distributed with respect to the cardinal points in the stem cross-section, even in a stand
situated on a slope [25]. It was observed that the occurrence of ring shake is
concentrated in a limited number of rings which are characterised by a narrow radial
current increment followed by larger rings, i.e. in trees that have grown irregularly [4,
17, 22, 25, 54].
34
Section B
Towards the causes
We can define the cause of a specific defect as the antecedent event, condition, or
characteristic that is necessary for the defect to occur at the moment that it did [60]. The
concept of causation is commonly characterised by the assumption that there is a oneto-one relationship between the observed cause and the effect. But with experience and
research into the process that causes ring shake, we have been persuaded that ring shake
results from a complexity of factors that act in concert. If we consider the formation of
ring shake from a mechanical point of view, the fracture appears in the wood when the
radial strength is weaker at a given time and in a given place than the stress acting in
that direction. From this standpoint, strength and stress are the key players in the
development of a rupture. Thus, we can analyse the formation of ring shake by focusing
our attention on the equilibrium between these two central factors.
Weak radial wood strength
Chestnut is widely known to be a very fissile wood. Its strength perpendicular to the
fibre is almost half that of oak (Table II) [12]. Many studies of the transversal
mechanical strength of chestnut have shown that trunks with ring shake display a lower
average radial strength value than those without ring shake [4, 20, 21, 32, 51-53, 63].
This kind of evidence, however, does not explain the entire phenomenon. In fact the
experiments carried out did not establish any statistical value of performance to
distinguish ring shaken trees from unshaken ones. It was also observed that wood
strength distribution is not homogeneous along the radius. Results from various studies
indicate that radial wood strength decreases from the pith to the bark [32, 53], probably
as a result of decreasing specific density from the pith to the bark, which may be in
relationship with the radial wood strength.
Table II. Ring shake-relevant characteristics of chestnut wood
Direction
Longitudinal
Tangential
Radial
Source
E-modulus
σstrength
[MPa]
[MPa]
1400-2400 * 100
[51]
7-16 * 100
[51]
Instantaneous deformations
on stem surface
[%]
a
- 0.095
b
0.110
b
- 0.059
a
[63]
b
[43]
Hygrothermal
deformations
[%]
Drying
deformations
[%]
± 0.1
From 0.4 to 0.6
- 0.1
[9, 38, 42]
-0.33
-8.08
-3.44
[51]
Individual tree effect (Genetic factors?)
The fact that chestnut wood is extremely weak in some cases might be primarily due to
genetic causes. Two main observations lead us to suppose that genetic factors regulate
wood strength, and so indirectly ring shake formation. First, ring shake is not randomly
distributed within a single stand, but is instead concentrated over several trees (stools)
that are particularly prone to this defect. Second, fracture tests performed with different
trees grown in the same stand revealed that strength is principally an intrinsic
characteristic of the individual [20, 51]. In particular, it is common for all the shoots of
the same coppice stump to behave in the same way [25, 54]. These observations suggest
that a link exists between genetic factors and ring shake formation. This hypothesis is
supported, although not proven, by results from various studies aiming to establish a
link between radial strength and genetic factors [30-32, 63].
Section B
35
Soil effect?
It has been observed that radial strength is lower in those stands where the soil is
particularly poor in calcium and other cations [63]. In fact, it is claimed that the binding
capacity of calcium cations can strengthen the middle lamellas, as described by the
"egg-box model" developed by Grant et al. [37]. Several studies have been performed to
clarify the relationship between calcium contents in the soil and in trees, with a special
regard to ring shake incidence [33, 34, 49, 50, 59, 63, 66, 67]. The results obtained
indicate a link between very low calcium contents and ring shake incidence, yet without
providing the evidence for a direct relationship between them. We must also underline
that as a species, chestnut is known to be intolerant of calcareous soils: it is possible that
the problem is due to difficulties in calcium absorption, rather than the absolute amount
of calcium.
The stresses in wood
Apart from the external and temporary stresses that may act on trees and installed wood,
such as wind and snow, three mechanisms could be responsible for the stresses that
cause splitting: the instantaneous release of certain growth stresses as a result of treefelling, stem-crosscutting and log-sawing [5, 47]; the additional relieving of stresses that
is observed when wood is heated (hygrothermal recovery) [38, 40, 45, 46]; and the
stresses generated as a consequence of anisotropical shrinkage of wood. Both
instantaneous stress release and hygrothermal recovery seem to be related to the
rheological conditions of wood cell maturation and of morphological tree growth [39,
64], while drying stress originates in the moisture change process in wood and is linked
to the drying process parameters.
Growth stresses
The term "growth stress" refers to the distribution of mechanical stresses that develop in
stems as the tree grows in diameter and height. This is the result of the superposition of
support stresses and maturation stresses [27, 28]. Support stresses are caused by the self
weight supported by the tree; their distribution in the stem depends heavily on the
historical evolution of the tree-loading and on the existing geometrical and architectural
situation [27] and are difficult to estimate. More relevant to the formation of ring shake
are maturation stresses. This kind of stress arises during the maturation of new cells.
Just after differentiation, as cells mature, they are subjected to bio-mechanical
transformations that occur at the S2 cell wall level [7, 13]. As a result, cells tend to
modify their dimensions and generate stresses in wood. Several authors have proposed
models describing the distribution of maturation stresses within the stem [5, 6, 26, 28,
47]. Using these models, Thibaut et al. [63] gave a qualitative illustration of how
longitudinal, tangential and radial stresses are distributed in the stem (Figure 2). Near
the bark there is longitudinal tension, tangential compression and no radial stress, while
near the pith there should be a high level of longitudinal compression and both radial
and tangential tension. Wood is mostly prone to break in tangential or radial tension. So
end splitting linked to growth stress relief must occur near the pith [47] and should take
the form of radial cracks. This is the case even in chestnut for which has always had at
least a small crack running from the pith outwards [57]. This also suggests that no ring
shake should occur near the bark, as was observed [11]. After the first small radial crack
occurs, stress distribution inside the log is changed and maximum radial stress is then
located between the end of the crack extension and the middle of the radius [9, 38]. This
36
Section B
probably explains the numerous observations of ring shake distribution along the radius
cited before.
Chestnut coppice shoots reveal longitudinal surface strain stress values (Table II)
similar to other broad-leaved species like beech, eucalyptus and poplar [29]. No
substantial interregional differences in measured stress have been observed and coppice
management does not seem to favour higher values except for at the base of curved
trunks [63]. In symmetrical stools in fact, longitudinal deformation is usually constant
along the circumference of the shoots, the value being characteristic of each single [65].
Some stools however displayed stem's sector with longitudinal deformation 5 times
greater than the "standard" values [21, 29, 65]. This phenomenon comes back to the
heterogeneous distribution of reaction wood, which possesses a particularly high
maturation stress, in the stem. The same authors also observed that sectors characterised
by small annual rings have a lower longitudinal surface deformation value than sectors
with large annual rings. Although differences were observed between trees, no clear
relationship between high longitudinal surface strain and ring shake occurrence was
found. The same statement is also true for the transverse stresses measured on the same
sample of trees [63, 65].
8
7
6
Stress (MPa)
5
4
R, initial
T, initial T
R, central hole
T, central hole
R, crosscut
T, crosscut
3
2
1
0
-1 0
1
-2
Relative radius (r/R)
Figure 2. Transverse growth stresses at log ends before [42] and after crosscutting, and after the
appearance of small heart cracks. The model assumes an axisymmetric, homogeneous and transversally
isotropic log, constant maturation stress (Kubler's model [47]) and the heart cracks are made equivalent to
a central hole [40] (here 5% of log diameter).
Hygrothermal recovery
Locked strains in trees are partially released by cutting specimens from the tree, and
more completely through hygrothermal recovery, by boiling them in a green state, so as
to exceed the softening point of lignin [39, 45, 46]. Hygrothermal recovery evidences
the effect of the transverse strains [42] (Table II). As a result hygrothermal recovery
causes the further growth or new development of either radial cracks in some trees or of
ring shakes in others [57, 63]. This indirectly proves that residual stresses that are
Section B
37
distributed like growth stresses are prone to develop ring shake and that there are two
populations of logs: those that extend heart checks without ring shakes and those that
extend ring shakes leaving the first heart shakes at their low extension.
Drying stresses
During drying, timber moisture content decreases from a very high level to a relatively
low level. When the bounded water of the cell walls is removed too, wood volume
begins to change an anisotropically, thereby generating drying stresses. Compared to
other similar species, chestnut wood does not display any anomalous shrinkage (Table
II) or unusual microfibril angle [62], which might justify its high occurrence of ring
shake. In his study, however, Leban [51] observed that within the same radial section,
ring shaken annual rings display a double radial shrinkage compared to the mean radial
shrinkage of the whole radius. In addition, tangential shrinkage was always found to be
higher in the annual ring preceding the ring shaken one. Fioravanti [23] also noticed
different shrinkage in wood: he observed that there is a different longitudinal shrinkage
in the earlywood and latewood areas of the same annual ring. This gradient was
particularly pronounced in rings characterised by a small annual increment surrounded
by larger rings. All these observations lead us to assume that the structure of the single
annual ring may influence local shrinkage and consequently could further favour the
development of ring shake in specific annual rings.
Concluding remarks
The phenomenon of ring shake appears with a high level of variability, making
comprehension of the process leading to ring shake a hard task. However, by combining
consistent results from different studies, we are able to draw up a possible scenario for
ring shake formation. The key elements are radial wood strength and wood stresses.
From a mechanical point of view, the mechanism that induces ring shake is simple:
tangential separation occurs if radial wood strength is weaker than the wood stress
acting in that direction. It is more difficult to prove where and when this condition is
achieved. Both those elements are regulated by several factors as described in figure 3.
It has been known for a long time that traumatic ring shake is trauma-related in origin
and that this type of fracture is not the most important facet of the chestnut ring shake
problem because we know causes and possible remedies [18]. The situation with
healthy-type ring shake is more problematic, however. On the basis of wha t we have
described above, we can assert that the phenomenon of healthy ring shake in chestnut is
principally linked to the weak wood strength of this species. It is striking that chestnut
wood tends to develop tangential splits while most of the other wood species form radial
fractures. This particular behaviour of chestnut wood and the fact that it is the wood
most commonly affected by ring shake suggest that in the radial direction this type of
timber might be particularly weak compared to other species. Ferrand [22] hypothesised
that this weakness might be brought back to the singularities in the structure of chestnut
wood (Figure 4).
38
Section B
Heritable
information
Wood
properties
Environmental
effect and growth
history
Wood
processing
Growth
stresses
Drying
stresses
Wood
stresses
σstrength < σstresses è Ring shake
Figure 3. Diagram of ring shake formation process. The key elements in ring shake formation are
radial wood strength (σstrength) and wood stresses(σstresses). Radial wood strength results from the
interaction between the anatomical, chemical and physical characteristics of wood, which are
determined by genetic constitution, tree environment and tree history. The same is true as for wood
stresses. In addition, wood stresses depend on the stage of wood processing, whereby stresses can be
relieved (growth stresses) or newly generated (hygrothermal recovery and drying stresses). We draw
attention to the fact that wood properties change in the stem as trees grow in size and height, thereby
establishing a dynamic and complex relationship between the aforementioned elements playing a role
in the development of ring shake.
Figure 4. Structure of chestnut wood. Electronic scanning microscope image. Characteristic of chestnut
wood are the ring porous wood structure with large earlywood vessels and the thin monoseriate radial
rays.
Two features in particular are characteristic of chestnut wood anatomy. It displays a
ring-porous wood structure, in which the earlywood vessels are distinctly larger than the
latewood ones, generating a soft zone rich in cavities and a very distinct and
homogeneous interface between successive rings. Second, being of monoseriate type,
radial rays that act as radial reinforcing fibres [2, 3, 8, 14, 44, 55, 56] can only partially
Section B
39
fulfil this function in chestnut wood. Thus, it is easy for tangential cracks to propagate.
In contrast, the radial cracks that usually form along radial rays have to cross both the
soft earlywood and the rigid latewood zones, which offer resistance to fracture
propagation. Depending on the balance between radial and tangential strength, will it
develop either ring shake or radial cracks. Using the results of studies conducted to date,
it is possible to suggest that genetic constitution and soil nutrient content are key
determinants of a tree's susceptibility to ring shake and thus determine whether it will
tend to form radial cracks or ring shake. But it is not yet quite clear what links exist
between genetics or nutrients and chestnut wood microstructure, both at the cellular
level (geometry of vessels, fibres or rays for example) and at the cellular wall level
(compound middle lamella architecture) that could explain this susceptibility to ring
shake. At all events, ring shake will also occur mainly where stresses are worse for this
kind of rupture. This explains the higher probability of ring shake near the inner third of
the stem radius and at the bottom of the felled stem (felling stresses, and first stress
recovery in cross-cutting). It is probably also the reason why there is a correlation
between irregular diametric growth (i.e. high heterogeneity in ring width) and ring
shake occurrence. This should lead to high levels of local stress linked to heterogeneity
in maturation stresses combined with local changes of wood properties (shrinkage for
example), both just after harvesting, or during wood processing. There is also evidence
that the older a chestnut tree is, the higher the probability of ring shake occurrence will
be. It is not quite clear if this is a consequence of ageing on wood strength, of dimension
on growth stress distribution or simply a mechanical effect of the growing probability of
irregular growth with passing time, where occasional very dry or cold seasons are
responsible for narrow rings.
Considering the aforementioned aspects, and in order to limit the follow-on damage of
ring shake, the following could be taken into account wherever quality chestnut wood
(with a lower risk of ring shake) has to be produced. There are three points in the wood
production process where decision can be made that affect the likelihood of ring shake.
The first has to be made when selecting a site. Although chestnut does not like
calcareous soils, it is better to grow chestnut trees on fertile soils where, in particular,
there is enough free calcium. Not only because calcium may reduce the risk of ring
shake, but also because a fast and regular growth helps to reduce the risk of ring shake.
Coppice stands are to be preferred to high forest because of the short rotation time, but
active silviculture is required in order to maintain regular growth and an equilibrated
tree shape. The second decision affects individual trees, and involves recognising their
genetic constitution as regards radial wood strength so as to eliminate the trees that are
prone to ring shake. At present, there are two possible techniques that might help, but
both need improvement. The first one is measuring radial strength on wood samples
taken directly from the standing trees, e.g. using Fractometer [36] tests. The second
method is measuring the radial propagation of ultrasound waves in the stem [35]; there
seems to be some relationship between radial wave propagation and the occurrence of
ring shake. The third decision is made when the tree is felled. Depending on the kind of
fractures observed at the basis of the stem (ring shake or radial crack), it is possible to
make a further selection and decide on the further industrial use of the wood. While
these measures do not completely exclude the possibility of ring shake, they certainly
minimise both the risk and the consequences of ring shake.
40
Section B
Future prospects
As we have seen, some aspects of the whole mechanism have yet to be explained and
certain relationships have not yet been completely demonstrated. Below we have listed
certain points that merit further investigation in order to obtain a better picture of the
complex phenomenon that is ring shake:
• different features of ring shake have been recognised. It is likely that the mechanism
that leads to each fracture feature has a different origin. A detailed description of this
aspect may help in comprehending the mechanism that causes breaking, in particular the
type of stress that is principally involved in the development of the fracture as well as
the characteristics of the broken material;
• the structure of chestnut wood is conducive to weak wood strength. A better
understanding of the relationships between ring shake incidence, anatomical and
mechanical wood characteristics and the influence of factors such as genetic
constitution and soil may help us better evaluate the risk of ring shake;
• irregularity in chestnut wood seems to influence ring shake. A better description of
this effect on wood properties and stresses is needed and should be further investigated.
An understanding of irregularity could be beneficial in developing silvicultural model
contributing to minimise the risk of ring shake development.
Acknowledgements
We sincerely thank Marco Conedera and Fulvio Giudici for their helpful comments and
are grateful to Joseph Gril for reviewing the paper.
Literature
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
Alberti, M., La cipollatura del castagno: ricerche sperimentali circa la sua localizzazione all'interno
dei singoli fusti, Tesi di laurea, Università degli studi di Torino, 1991. pp. 94.
Albrecht, W., Untersuchung der Spannungssteuerung radialer Festigkeitsverteilung in Bäumen,
Dissertation, Universität Karlsruhe, 1995. pp. 121.
Albrecht, W. A., Bethge, K. A.Mattheck, C., Is lateral strength in trees controlled by lateral
mechanical stress?, Journal of Arboriculture 21 (1995) 83-87.
Amorini, E., Bruschini, S., Fioravanti, M., Macchioni, N., Manetti, M. C., Thibaut, B.Uzielli, L.,
Studi sulle cause di insorgenza della cipollatura nel legno di castagno (Castanea sativa Mill.), in:
Comunità montana delle Prealpi Trevigiane (ed), Atti del Convegno nazionale sul castagno, Cison
di Valmarino (Treviso), 23-25 ottobre 1997, 1998, pp. 269-292.
Archer, R., Growth stresses and strain in trees, Springer Verlag Berlin, 1986.
Archer, R.Byrnes, F. E., On the distribution of tree growth stresses. Part I: An anisotropic plane
strain theory, Wood Sci. Technol. 8 (1974) 184-196.
Bamber, R. K., The origin of growth stresses: a rebuttal, IAWA Bul. 8 (1987) 80-84.
Beery, W. H., Quantitative wood anatomy: relating anatomy to transverse tensile strength, Wood
Fiber Sci. 15 (1983) 395-407.
Berrada, E., Recouvrance hygro-thermique du bois vert, Thèse, Université des sciences et
techniques du Languedoc Montpellier, 1991. pp. 281.
Boetto, G., La cipollatura del castagno: osservazioni in due boschi cedui del Piemonte (Chiusa
Pesio CN, Bibiana TO), Tesi di laurea, Università degli studi di Torino, 1991. pp. 104.
Bonenfant, M., Croissance et qualité du châtaignier de futaie en Bretagne, Mémoire, SERFOB
Bretagne, 1985. pp. 123.
Bourgeois, C., Le châtaignier: un arbre, un bois, Institut pour le développement forestier IDF,
Paris, 1992.
Section B
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
41
Boyd, J. D., Tree growth stresses. Part V: evidence of an origin in differentiation and lignification,
Wood Sci. Technol. 6 (1972) 251-262.
Burgert, I., Die mechanische Bedeutung der Holzstrahlen im lebenden Baum, Dissertation,
Universität Hamburg, 2000. pp. 173.
Chang, C. I. J., The cause of ring shake: a review of literature, Quarterly Journal of Chinese
Forestry 6 (1972) 69-77.
Chanson, B., Approche structurale et histophisiologique du bois de châtaignier (Castanea sativa
Mill). Etude préliminaire à celle de la roulure, D.E.A. Sciences forestières, Université de Nancy 1,
1982. pp. 34.
Chanson, B., Etude de la variabilité de quelques propriétés physique et anatomique du bois de
châtaignier (Castanea sativa Mill). Application à l'étude de la roulure, Thèse, Université des
sciences et techniques du Languedoc Montpellier, 1988. pp. 278.
Chanson, B., Leban, J. M.Thibaut, B., La roulure du châtaignier (Castanea sativa Mill.), For. Med.
11 (1989) 15-32.
Cielo, P., Incidenza e tipologie della cipollatura in un ceduo di Castagno del Comune di Garessio
(CN): ricerche bibliografiche e sperimentali, Tesi di laurea, Università degli studi di Torino, 1988.
pp. 356.
Dumonceaud, O., Etude de la fiabilité du fractomètre de Claus Mattheck en vue de son utilisation
comme appareil de détection du risque d'apparition de la roulure chez le châtaignier (Castanea
sativa Mill.), Mémoire, Université des sciences et techniques du Languedoc Montpellier, 1994. pp.
80.
Elzière, S., Influence de l'éclaircie sur des brins de taillis de châtaignier (Castanea sativa Mill.)
Analyse de tige, contraintes de croissance et fissilité du bois, Mémoire, Ecole national d'ingénieur
des travaux agricoles de Bordeaux, 1995. pp. 57.
Ferrand, J. C., La roulure du châtaignier: Note préliminaire, INRA - CNRF Station de recherche
sur la qualité du bois, 1980. pp. 15.
Fioravanti, M., Caratterizzazione del legno giovanile di castagno (Castanea sativa Mill.): studi su
anatomia, densitometria e variazioni dimensionali, Tesi di laurea, Università degli studi di Firenze,
1992. pp. 75.
Fonti, P., Studio delle correlazioni tra alcune caratteristiche del soprassuolo e dei polloni di
castagno (Castanea sativa Mill.) con l'incidenza e la distribuzione della cipollatura in un ceduo a
Novaggio, Ticino, Lavoro di diploma, Politecnico federale di Zurigo, 1997. pp. 88.
Fonti, P., Giudici, F., Kucera, L. J., Ott, E.Pöhler, E., Studio sulla cipollatura in un ceduo castanile,
in: Comunità montana delle Prealpi Trevigiane (ed), Atti del Convegno nazionale sul castagno,
Cison di Valmarino (Treviso), 23-25 ottobre 1997, 1998, pp. 293-302.
Fournier, M., Bordonne, P. A.Guitard, D., Growth stress patterns in tree stems: A model assuming
evolution with the tree age of maturation strains, Wood Sci. Technol. 24 (1990) 131-142.
Fournier, M., Chanson, B., Thibaut, B.Guitard, D., Mécanique de l'arbre sur pied: modélisation
d'une structure en croissance soumise à des chargements permanents et évolutifs. 1. Analyse des
contraintes de support, Ann. sci. for. 48 (1991) 513-525.
Fournier, M., Chanson, B., Thibaut, B.Guitard, D., Mécanique de l'arbre sur pied: modélisation
d'une structure en croissance soumise à des chargements permanents et évolutifs. 2. Analyse
tridimensionnelle des contraintes de maturation, cas du feuillu standard, Ann. sci. for. 48 (1991)
527-546.
Fournier, M., Chanson, B., Thibaut, B.Guitard, D., Mesures de déformation résiduelles de
croissance à la surface des arbres en relation avec leur morphologie. Observations sur différentes
espèces, Ann. sci. for. 51 (1994) 249-266.
Frascaria, N., Etude de la diversité génétique du châtaignier (Castanea sativa Mill.) en île de
France: essai de corrélation avec des marqueurs morphologique, phénologique et de qualité du
bois (roulure), D.E.A. d'Ecologie générale, Université de Paris-Sud, 1987. pp. 34.
Frascaria, N., Variabilité génétique d'un arbre forestier: le châtaignier (Castanea sativa Mill.),
Thèse, Université de Paris-Sud, 1991. pp. 122.
42
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
Section B
Frascaria, N., Chanson, B., Thibaut, B.Lefranc, M., Génotypes et résistance mécanique radiale du
bois de châtaignier (Castanea sativa Mill.): Analyse d'un des facteurs explicatifs de la roulure,
Ann. sci. for. 49 (1992) 49-62.
Freyssac, C., Laroche, A., Carlue, M.Morvan, H., Contribution à l'étude de la roulure chez le
châtaignier: étude expérimentale des conséquences d'amendements calciques. II - Effet du calcium
sur le contenu pectique du bois, Annales Scientifiques du Limousin 10 (1994) 35-43.
Freyssac, C., Rahmani, A., Carlue, M., Verger, J. P.Morvan, H., Contribution à l'étude de la
roulure chez le châtaignier: étude expérimentale des conséquences d'amendements calciques. I Relation entre la composition minérale du milieu et celle de jeunes plants, Annales Scientifiques
du Limousin 10 (1994) 25-34.
Giudici, F., Fonti, P., Pöhler, E. Sandoz, J. L., Qualità del legname di castagno: diagnosi della
cipollatura per mezzo di ultrasuoni, in: Comunità montana delle Prealpi Trevigiane (ed), Atti del
Convegno nazionale sul castagno, Cison di Valmarino (Treviso), 23-25 ottobre 1997, 1998, pp.
259-267.
Götz, K. O.Mattheck, C., Studies on the strength of green trees using the fractometer III: model
description and user's manual, Wissenschaftlichen Berichte, Forschungszentrum Karlsruhe, 1999.
pp. 15.
Grant, G. T., Morris, E. R., Rees, D. A., Smith, P. J. C.Thom, D., Biological interaction between
polysaccharides and divalent cations: the egg-box model, FEBS Letters 32 (1973) 195-198.
Gril, J., Berrada, E.Thibaut, B., Recouvrance hygrothermique du bois vert. II - Variations dans le
plan transverse chez le châtaignier et l'épicéa et modélisation de la fissuration à coeur provoquée
par l'étuvage, Ann. sci. for. 50 (1993) 487-508.
Gril, J.Thibaut, B., Tree mechanics and wood mechanics: relating hygrothermal recovery of green
wood to the maturation process, Ann. sci. for. 51 (1994) 329-338.
Gril, J., Thibaut, B., Berrada, E.Martin, G., Recouvrance hygrothermique du bois vert. I Influence de la température: Cas du jujubier (Ziziphus lotus (I) Lam.), Ann. sci. for. 50 (1993) 5770.
Guiot, A., Contribution à l'étude du châtaignier à bois en Bretagne, SRAF de Bretagne, 1983. pp.
37.
Jullien, D., Analyse expérimentale et numérique des contraintes résiduelles dans un matériau
élastique orthotrope élaboré par couches successives. Cas d'un disque de bois vert., Thèse,
Université des sciences et techniques du Languedoc Montpellier, 1995. pp. 214.
Jullien, D.Gril, J., Mesures des déformations bloquées dans un disque de bois vert. Méthode de la
fermeture, Ann. sci. for. 53 (1996) 955-966.
Keller, R.Thiercelin, F., Influence des gros rayons ligneux sur quelques propriétés du bois de
hêtre, Ann. sci. for. 32 (1975) 113-129.
Kubler, H., Hygrothermal recovery under stress and release of inelastic strain, Wood Sci. 6 (1973)
78-86.
Kubler, H., Role of moisture in hygrothermal recovery, Wood Sci. 5 (1973) 198-204.
Kubler, H., Growth stresses in trees and related wood properties, For. prod. abstr. 10 (1987) 61119.
Lachaussée, E., Note upon shake and forest crack of Quercus robur, For. Abstr. 15 (1953) 1598.
Laroche, A., Approche expérimentales de la nutrition du châtaignier: Influence d'apports nutritifs
sur la croissance des plants, sur la distribution des éléments minéraux dans les organes et sur les
caractéristiques générales des pectines, chez des jeunes arbres cultivés sous serre, Thèse,
Université de Limoges, 1997. pp. 188.
Laroche, A., Freyssac, V., Rahmani, A., Verger, J. P.Morvan, H., Growth and mineral content of
chestnut trees under controlled conditions of nutrition, Ann. sci. for. 54 (1997) 681-693.
Leban, J. M., Contribution à l'étude de la roulure du châtaignier, Thèse, Institut National
Polytechnique de Lorraine, 1985. pp. 164.
Macchioni, N., Studi sulla cipollatura del castagno (Castanea sativa Mill.): metodologie per la
valutazione della coesione trasversale del legno, Te si di Laurea, Università degli studi di Torino,
1992. pp. 102.
Section B
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
43
Macchioni, N., Mechanical strength and ring shake in chestnut (Castanea sativa Mill.), For. Med.
16 (1995) 67-73.
Macchioni, N.Pividori, M., Ring shake and structural characteristics of a chestnut (Castanea sativa
Mill.) coppice stand in northern Piedmont (Northwest Italy), Ann. sci. for. 53 (1996) 31-50.
Mattheck, C., Albrecht, W.Dietrich, F., Die Biomechanik der Holzstrahlen, Allg. Forst- Jagdztg.
165- (1994) 143-147.
Mattheck, C.Schwarze, M. R., Die Holzstrahlen als getarnte I-Balken in einem mechanischen
Ersatzmodell für Holz, Allg. Forst- Jagdztg. 165 (1994) 197-201.
Movassaghi, E., Mothe, F.Thibaut, B., Etude de la faisabilité du déroulage de brins de taillis de
Châtaignier, Laboratoire de mécanique générale des milieux continus, Université de Montpellier,
1987. pp. 50.
Pozzi, F., Studio anatomico comparativo sulle modalità di frattura del legno di castagno (Castanea
sativa Mill.): campioni affetti da cipollatur a e campioni rotti meccanicamente, Tesi di laurea,
Università degli studi di Torino, 1996. pp. 90.
Ranger, J., Felix, C., Bouchon, J., Nys, C.Ravart, M., Dynamique d'incorporation du carbone et
des éléments nutritifs dans un taillis simple de châtaignier (Castanea sativa Mill.), Ann. sci. for. 47
(1990) 413-433.
Rothman, J. K.Greenland, S., Modern epidemiology, Lippincott-Raven Philadelphia, 1998.
Saya, I., Indagini sulla cipollatura del legno di Castagno e di Abete bianco, in: Contributi
scientifico-pratici per una migliore conoscenza e utilizzazione del legno, CNR, I. p. l. r. s. l.,
Firenze, 1962, pp. 37-41.
Staglianò, N., Caratterizzazione ultrastrutturale del legno di castagno: studi sull'angolo di
inclinazione delle microfibrille, Tesi di laurea, Università degli studi di Firenze, 1992. pp. 75.
Thibaut, B., Chanson, B., Fournier, M.Jullien, D., Valorisation du bois de châtaignier: prévoir et
réduire les risques de roulure à la production, Rapport Final, Institut pour le Développement
Forestier, 1995. pp. 48.
Thibaut, B., Fournier, M.Jullien, D., Contraintes de croissance, recouvrance différée à l'étuvage et
fissuration des grumes: cas du châtaignier, For. Med. 16 (1995) 85-91.
Uzielli, L., New silvicultural methods and innovative technologies for the valorisation of Chestnut
wood as a prime resource for industry, EEC Research Program, Final report, Task C, Operative
Unit 06, Forest Program Chestnut - Task C, 1994. pp. 38.
Verger, J. P., Domain, P., Fournier, J. M., Maisonier, C.Djomo, J. E., Importance du calcium dans
le développement in situ du taillis de châtaignier en Limousin, Ann. sci. for. 10 (1994) 13-24.
Verger, J. P., Morvan, H., Fournier, J. M., Matga, F., Maisonier, C., Freyssac, V., Desjobert,
T.Domain, P., Nutrition minérale du châtaignier (Castanea sativa Mill.): rôle dans le
développement de la roulure, 1993. pp. 36.
Section B
45
2
Ring shake in chestnut: anatomical description,
extent and frequency of failures
Patrick Fontia and Nicola Macchionib
a
Via Belsoggiorno 22, Casella postale 57, CH-6504 Bellinzona, Switzerland
b
Istituto per la ricerca sul legno, Consiglio nazionale delle ricerche,
Via A. Barazzuoli 23, I-50136 Firenze, Italy
Annals of Forest Science (submitted 2002)
Submitted: 27 March 2002
46
Section B
Abstract
Ring shake is a wood defect that occurs very frequently in sweet chestnut (Castanea
sativa Mill.). By examining this particular kind of wood failure it is apparent that, at an
anatomical level, separations occurring in the wood tissue show different features. In
order to collect further information to help in understanding the causes that lead to the
development of ring shake, a microscopic description and a quantification of these
different ring shake typologies has been performed on 45 chestnut wood discs. Results
showed that among the various shake types encountered, the crack-form, i.e. the failure
that develops across the earlywood cell walls, is largely the most frequent and is
principally found in fresh felled wood. Detailed observations reinforce the belief that
the set off mechanism leading to crack-failure is related to the combined interaction of
the structural weakness of chestnut wood with growth stresses developed in the stem.
Castanea sativa / ring shake / fracture anatomy.
Résumé
La roulure du châtaignier: description anatomique, ampleur et fréquences des
séparations - La roulure est un défaut du bois qui se produit très fréquemment dans le
châtaignier (Castanea sativa Mill.). En examinant ce genre particulier de rupture il est
visible que, au niveau anatomique, les séparations qui se produisent dans le tissu bois
montrent des morphologies de fracture différentes. Afin de disposer davantage
d'informations nécessaires à la compréhension des causes qui mènent au développement
de la roulure, une description microscopique et une quantification de ces différentes
typologies de roulure a été effectué sur 45 disques en bois de châtaignier. Les résultats
ont montré que parmi les divers types de roulure observés, la forme de rupture ¨crack¨,
c.-à-d. la séparation qui se développe à travers les parois cellulaires des vaisseaux du
bois initial, est la plus fréquente et se présente principalement dans le bois fraîchement
abattu. Les observations microscopiques détaillées renforcent la conviction que le
mécanisme conduisant à la forme de rupture ¨crack¨ est lié à la faiblesse structurale du
bois de châtaignier combinée avec les contraintes de croissance développés dans la tige.
Castanea sativa / roulure / anatomie de fracture.
Section B
47
Introduction
Looking at the stem cross-section of different wood species we can observe the
presence of circular failures running parallel to the growth ring which strongly
downgrade otherwise valuable timber. This particular kind of tangential shake is
commonly called “ring shake” and is defined as “a lengthwise separation of wood
which occurs between and parallel to the growth layers”. The development of such
fractures depends on the balance between stresses and strength: when the first one
exceeds the second then the shake opens. The literature analysis reveals many different
opinions as to the causes of shake, but all either refer to an increase of stresses due to
the effect of wind [10, 26], frost [8, 10], sudden changes in diameter growth rates [12],
relieving of growth stresses [1, 13]; or to a weakened wood strength caused by cambial
damage [11, 16, 18, 21, 22, 25], environmental stress [25] or by a lack of substances in
the soil and the tree [14]. Several studies based on the description of the anatomical
features of shake surfaces permitted further discussion about the possible causes of ring
shake and about when they have likely occurred [16, 17, 25].
Castanea sativa is a hardwood species that is very often affected by ring shake [4]. In
this species different types of ring shake separations have been observed. Chanson et al.
[3], considering the origins of the split, distinguished between “traumatic” and
“healthy” ring shake. The first type is always related to visible anomalies in the wood
tissue, whereas for the second one the splitting appears to be unrelated to any
recognizable anatomical perturbation. Moreover, among these two types, three
additional anatomical fracture surfaces have also been observed: the “overlay”, that
corresponds to a new layer superposed on traumatic cells; the “detachment”, that
appears as a separation along the compound middle lamella between cells but leaving
intact the cell walls; and the “crack”, where the failure develops across the earlywood
vessel cell walls [7]. While it is clear that the overlay originates from cambium damage,
the reasons for the development of the other two features are rather unclear, even if all
the authors dealing with this species agree on the mechanical failure hypothesis. It is
however possible that the cracks and the detachment features have different origins or
developmental processes that lead to dissimilar anatomical failure characteristics. Thus
these aspects have to be taken into account when discussing the possible causes of ring
shake. To this date, however, no descriptive studies on the anatomical characteristics of
the ring shake zones in chestnut have been carried out, except for a limited study by
Saya [20].
The objective of this research is therefore to perform a detailed description of the
different anatomical ring shake fractures. In particular we aim to furnish a report on
wood surface separations from the anatomical point of view, as well as a quantification
of their extent and frequency. We hope therefore, through this new knowledge, to
contribute to the discussion of the possible causes leading to ring shake in chestnut
wood.
Material and methods
The wood material used for this study originates from 2 mature coppice stands situated
in Brione and Bedigliora, in the Southern part of the Swiss Alps. From each stand at
least 20 shoots were selected from the overstory trees, 24 from the stand in Brione and
21 from that of Bedigliora. Immediately after the felling, 5 cm thick wood discs were
gathered from the stem base (50 cm above ground level) of each selected shoot. The
Annals of forest science, submitted 2002
48
Section B
collected discs were then polished with a 150 grit sand paper in order to obtain a clean
cross-section allowing an easy identification and characterization of the anatomical
details of the different ring shake typologies. The observations were performed on each
wood disc twice: once on the fresh collected green wood discs, i.e. within 3 days after
the tree-felling (discs were stored in a controlled environment at 20°C and 90% relative
humidity), and then repeated on the same wood discs after being dried in fresh air under
shelter for about 1 year.
The characterization and quantification of ring shake was performed on the disc crosssections (Figure 1). All visible tangential failures longer than 1 cm were taken into
account. Each single ring shake was then characterized by its position on the disc
(distance from the pith, orientation and solar year of the annual ring affected by the
split), length and type of wood failure (overlay, detachment or crack).
Using a Scanning Electron Microscope (SEM Philips XL 20), a detailed microanatomical description of the surfaces of the different failure typologies was made on
the tangential and/or transversal plane of a few samples (two from each ring shake
typology) from the collected wood material. After boiling in water microscopic samples
were cut from the transversal and radial planes, without altering the tangential surface
failure caused by ring shake.
Detachment
crack
West
55 65 75 85 95
North
Brione 2:
green wood disc
East
Number of ring shake: 10
Total failure length: 160 cm
South
Figure 1: Schematic location of shakes on disc. The year of the annual ring is represented by dotted
lines, one each 10 rings starting from year 1955.
Observations
Anatomical features of shake zone
Overlay
The overlay-form is not exactly a fracture because a connection between the two
separated layers never subsisted. In fact, as a cicatrisation caused by cambial damage,
the tree from the near non-traumatic tissue just superposed, without any physical
connection, a new annual increment layer onto the traumatic one. Moreover, as a
consequence of micro organism penetrated in the wound, very often the nearby zone
display discoloration and decay. The overlay-form is therefore easy to recognise thanks
to the characteristic overgrown callous tissue and the discoloration in the nearby area
(Figure 2). The overlay-form is not exactly a fracture because a connection between the
Ring shake typologies
Section B
49
two separated layers never existed. In fact, as a cicatrisation caused by cambial damage,
the tree from the near non-traumatic tissue just superposed, without any physical
connection, a new annual increment layer onto the traumatic one. Moreover, as a
consequence of micro-organisms that penetrated the wound, the nearby zone often
displays discoloration and decay. The overlay-form is therefore easy to recognise thanks
to the characteristic overgrown callous tissue and the discoloration in the nearby area
(Figure 2). The injury causing the cicatrisation is a sudden event that acts on the last
formed tissue: its aspect is then random and its anatomical description cannot be
generalised.
Detachment
The fractures of the detachment typology are mainly located along the boundary
between two annual rings, even if from time to time a small excursion into the
earlywood area is noted, giving therefore the impression of a “crack” failure (Figure 3).
The failed surface is mainly smooth because the mechanism of tissue failure is a cell-tocell debonding from the compound middle lamella rather than a cell wall failure. As
figure 3 shows the failed surface does not exhibit any sign of cell cracking, even for the
parenchyma ray cells, whose walls at the end of the annual growth ring are still intact.
Here the ray cells are broken flush with the surface. A special case of detachment that
also rarely occurs is discoloured detachment, which is a disconnection that occurs due to
a weak bond between cells resulting from trauma and decay. Therefore, some
discoloration in proximity of the detachment is clearly visible.
Crack
From the anatomical point of view the crack-typology is essentially a cell wall failure
that mainly develops in the tangential plane crossing the first or second row of
earlywood vessels (Figure 4). The rough failed surface (tangential section) is
characterised by radial parenchyma cells (uniseriate rays) that still denote the pulling
effect that occurred during the opening of the failure, as well as by broken vessels and
disrupted fibres.
Figure 2: Overlay-form: the ring shake is due to the physical separation among the injured tissue and
the cicatrisation tissue.
50
Section B
Figure 3: Detachment-form: the failed surface is generally smooth even if a small excursion in the
earlywood area sometimes occurs (on the left image); the right image shows the aspect of the radial
parenchyma cells which are occluded between annual ring membranes.
Figure 4: Crack-form: the rough surface (on the left) is due to the fracture pattern running between the
earlywood vessels; the aspect of the ray tissue (on the right) is similar to that of “pulled” cells.
Extent and frequency of shakes
From the 45 selected wood discs, only a few were ring shake free. In fact only 8 (18%)
green discs and 4 (9%) dried discs showed no ring shake. Few other discs, 6 (13%)
green and 7 (16%) dried, were only slightly affected by ring shake, i.e. the total ring
shake failure length was less than 20 cm. This means that 69% of the green and 75% of
the dried discs displayed more than 20 cm of failure length. Among these, there were
two extremely shaken discs that exhibited more than 3 m of ring shake failure length.
As shown in Table I, the crack-form is largely the most frequent failure typology,
representing 88% (90% for the stand of Bedigliora and 86% for Brione) of all the ring
shakes observed in green wood, while in the dried wood this proportion slightly
diminishes to 84% (81% for Bedigliora and 86% for Brione). The other two failure
typologies are instead less recurrent with 11% (green) and 15% (dried) being of the
detachment typology and about 2% of the overlay-form, which of course did not
increase with the drying.
The green wood ring shakes represent 70% (3260 cm) of the entire observed ring
shakes, while a further 30% (1368 cm) developed during the drying process (Table 1).
These last formed fractures have been mainly surveyed on discs that were already
affected by the defect (Figure 5). The more the green wood is affected by ring shakes,
Section B
51
the more ring shakes, in particular of the crack typology, tend to develop during the
drying-process, confirming previous observations performed by Fonti et al. [6].
Table 1: Summary of the ring shake failure length observed in green and dried wood discs
differentiating between the different failure typologies. C=Crack; D = Detachment, O = Overlay, Σ = Sum
of all typologies
Ring shake failure
length
Bedigliora cm
%
Brione
cm
%
%
Total
cm
%
%
C
1205
90
1650
86
2855
88
Green wood discs
D
O
Σ
116
21
1342
9
1
100
229
39
1918
12
2
100
345
60
3260
11
2
100
Increment due to drying
C
D
O
Σ
451
252
703
64
36
0
100
583
81
655
88
12
0
100
1034 333
1368
76
24
0
100
C
1655
81
2233
84
3889
84
Dried wood discs
D
O
Σ
369
21
2045
18
1
100
310
39
2583
15
1
100
679
60
4628
15
1
100
Ring shake increment observed on the discs during the drying-prozess
140
Crack-form
Ring shake increment [cm']
Detach-form
Linear (Crack-form)
110
y = 0.2933x
R2 = -0.0925
80
50
20
-10 -
50
100
150
200
250
Total ring shake length observed on green wood discs (crack+detach) [cm']
Figure 5: Ring shake increment observed on each single disc during the drying process.
Shake distribution
The analysis of the ring shake distribution according to the year of the annual ring,
differentiated between the two stands, shows that both the detachment and crack-form
are not strictly related to specific annual increments (Figure 6). Looking however at
each single wood disc we often observed that failures, in particular the crack-form,
follow one ring for some distance but then rather abruptly cross radially into a
neighboring ring and then further proceed into the earlywood tissue. This “jumping”
from ring to ring could also occur several times in the same ring shake failure, giving
the failure a zigzag shape.
19
38
19
40
19
42
19
44
19
46
19
48
19
50
19
52
19
54
19
56
19
58
19
60
19
62
19
64
19
66
19
68
19
70
19
72
19
74
19
76
19
78
19
80
19
82
19
84
19
86
19
88
19
90
19
92
19
94
19
96
19
98
20
00
Average ring shake failure length [cm]
19
38
19
40
19
42
19
44
19
46
19
48
19
50
19
52
19
54
19
56
19
58
19
60
19
62
19
64
19
66
19
68
19
70
19
72
19
74
19
76
19
78
19
80
19
82
19
84
19
86
19
88
19
90
19
92
19
94
19
96
19
98
20
00
Average ring shake failure length [cm]
52
Section B
5
Bedigliora
4
Crack
Crack
Detach
Overlay
3
2
1
0
Year of the annual ring
14
Brione
12
10
Detach
Overlay
8
6
4
2
0
Figure 6: Ring shake distribution.
Section B
53
Discussion and conclusion
Origin
Several authors mainly associated ring shake formation with injuries [11, 16, 18, 21, 22,
25]. In this descriptive study on chestnut wood, ring shakes originating from an evident
traumatic event (overlay) represented only 1% of all the failure lengths observed.
The detachment-shake, i.e. failure arising in the compound middle lamella layer, is
somewhat more frequent (15%). In the literature this kind of fracture was often reported
as being located in the latewood area and associated with green wood ring shake [11,
15, 17]. In our study however the separation always appears at the ring boundary and
can be found both in green wood and as a result of the drying process. In her study,
Saya [20], possibly describing such a fracture, has observed a compound middle lamella
lignin deficiency. This could be indicative of external events that influenced the
bonding-quality between annual growth layers. Genetics, lack of substances in the soil
[14] or damage to the tree may be reasons for the poor bonding-quality. In fact, if the
shake is observed at a point far from the wound, the association of the wound with the
shake will not be directly recognised [22]. The rarely observed discoloured detachmentform supports this suggestion. And finally, as there is no apparent relationship between
specific annual increments and the occurrence of this kind of shake, the opening may
rather be associated with single tree events. However, cambial damage as a cause of
ring shake must be taken as a special case, and does not account for the majority of
shake.
The most frequent feature observed is indeed the crack-form, i.e. ruptures in which the
cell walls are usually broken, which represented 84% of the total ring shake failures
surveyed. In chestnut this separation principally develops in the earlywood area
proceeding across the large earlywood vessels. The fracturing that occurs appears to
involve a minimal quantity of wood.
With no noticeable abnormal wood tissue in the shake surface, this shake-form has
therefore a merely mechanical origin, which is regulated by the interaction between the
strengths and stresses in the wood.
The ring porous wood structure and the small uniseriate rays in chestnut confer a
privileged (weak) plane for the opening and propagation of this kind of tangential
failure. The earlywood zone is an area of structural weakness and therefore especially
vulnerable to stresses which could cause separation. The timing in which fractures
occur, which is generally before timber processing, and the often observed ring-to-ring
jumping pattern of shakes lets us presume that the set off mechanism leading to this
failure has to be related to stresses acting on the standing tree or on the fresh felled
stem.
These observations highlight therefore former hypotheses that the mechanism leading to
crack separation is related to the interaction between wood fragility and the release of
wood growth stresses [1, 2, 9, 13, 24]. The increase of new ring shake during the drying
process, limited to already severely shaken green wood discs, could therefore indicate
that each single disc has its own susceptibility to ring shake, which continues on into the
drying process. This may lead back to an intrinsic fragility or to an amount of
unrelieved stress specific to the considered individual.
54
Section B
Hypothesis on the mode of loading responsible for the different ring
shake forms
Based on the performed morphological description of the surface failures a preliminary
supposition about the responsible mode of fracture can be formulated. As opposed to
traumatic ring shake, healthy ring shake has a mechanical origin caused by the relieving
of growth stresses. However from simple fracture morphological observation it is
difficult to advance a hypothesis on the typologies of stresses (mode I, II or III) imposed
on the transversal/radial wood plane, even if some indications suggest allocating the
crack typology to mode I and the detachment typology to mode II. No studies of ring
shake in chestnut wood have compared the aspect of the ring shake failures to those of
the typologies of stresses imposed on the transversal-radial plane of wood. But
modelling the mode I and II crack propagation on spruce (Picea abies Karst.) wood,
Tan et al. [23] analysed by means of SEM the longitudinal tangential surface fractures.
Taking into account the differences between spruce and chestnut wood, we can observe
that the mode I produces fracture surfaces located within the earlywood, breaking the
cell walls, resulting in an aspect similar to what we have called “crack”. In spruce wood
the mode II produces a large number of warped broken tracheids, while the behaviour of
chestnut fibres is different and the mode I fracture causes the detachment of adjacent
fibres along the compound middle lamella.
An unpublished thesis from Pozzi [19] attempted an anatomical comparison between
the crack- and detach-failure ring shake typologies and those of samples artificially
broken in the tangential plane through radial bending and radial shear. This study
revealed that samples broken through radial bending performed on green wood (mode I)
are similar to the “crack-form” with the fissure breaking the thin and weak cell walls of
the first and second rows of earlywood vessels (Figure 7), while the shear fractures in
dried wood samples are similar to the “detachment-form” (Figure 8). It is therefore
reasonable to suppose that each typology of ring shake has its own mechanism. Further
investigation may help in establishing which fracture mechanism is responsible for
which kind of failure.
The anatomical observation suggests therefore that the development of the crack
typology seems to be linked to a radial bending stress that breaks the earlywood vessels.
These observations reinforce the hypothesis that ring shake develops frequently in
chestnut wood, because of the structural weakness that predisposes this species to this
kind of fracture, which is mainly regulated by the relieving of growth stresses during the
tree-felling. Therefore further studies on this topic, with the aim of reducing the risk of
ring shake, must consider in particular the development of the crack-form ring shakes
that appear before or immediately after the tree-felling. The few detachment failures
seem instead to be linked to the rolling shear stresses caused in the tangential plane by
the drying process along the abrupt transition between latewood and earlywood. At this
location there is an elevated difference in longitudinal shrinkage between the porous
zone and the preceding latewood [5], especially in the more fragile wood of already
shaken trees.
Section B
55
Figure 7: Facet of chestnut wood broken by radial bending test: the fracture line runs along the
earlywood vessels as in the “crack” form (Image Pozzi [19]).
Figure 8: Facet of shear fracture on chestnut wood: the fracture is similar to the detachment-form
(Image Pozzi [19]).
56
Section B
Acknowledgements
The Authors want to thank Simona Lazzeri, technician at the IRL – CNR, for preparing
the samples for SEM analysis and for the relevant images.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
Archer, R., Growth stresses and strain in trees, Springer Verlag Berlin, 1986.
Boyd, J. D., Tree growth stresses: II. The development of shakes and other visual failures in
timber., Aust. Jor. Appl. Sci. 1 (1950) 296-312.
Chanson, B., Leban, J.-M., Thibaut, B., La Roulure du châtaignier, Forêt méditerranéenne 11/1
(1989) 15-32.
Ferrand, J.-C., La roulure du châtaignier (Castanea sativa Mill.). Note préliminaire, Seichamps
(France CNRF) (1980) 17 p.
Fioravanti, M., Caratterizzazione del legno giovanile di castagno (Castanea sativa Mill.): studi su
anatomia, densitometria e variazioni dimensionali, Dottorato, Università degli studi di Firenze,
1992. pp. 75.
Fonti, P., Giudici, F., Kucera, L. J., Ott, E. Pöhler, E., Studio sulla cipollatura in un ceduo
castanile, Atti del convegno nazionale sul castagno, Cison di Valmarino (Treviso), 1998. 293-302.
Fonti, P., Macchioni, N., Thibaut, B., Ring shake in chestnut (Castanea sativa Mill.): state of the
art, Annals of Forest Science 59 (2002) 129-140.
Fostier, G., Frost injuries in Oak, For. Abstr. 14 (1953)
Fournier, M., Chanson, B., Thibaut, B., Guitard, D., Mécanique de l'arbre sur pied: modélisation
d'une structure en croissance soumise à des chargements permanents et évolutifs. 1. Analyse des
contraintes de support, Annals of Forest Science 48 (1991) 513-525.
Harrar, E. S., Defects in hardwood veneer logs: their frequency and importance. US Forest service
SE Forest experiment station 39 1954.
Kandeel, S. A.McGinnes, E. A. J., Ultrastructure of ring shake in scarlet oak (Quercus coccinea,
Muench), Wood science 2 (1970) 171-178.
Koehler, A., A new hypothesis as to the cause of shakes and rift cracks in green timber, Journal of
forestry 31 (1933) 551-556.
Kubler, H., Growth stresses in trees and related wood properties, Forest Products Abstracts 10
(1987) 61-119.
Lachaussée, E., Note upon shake and forest crack of Quercus robur, Forestry Abstracts 15 (1953)
1598.
McGinnes, E. A. J., Extent of shake in black walnut, Forest products journal 18 (1968) 80-82.
McGinnes, E. A. J., Phelps, J. E.Ward, J. C., Ultrastructure observations of Tangential shake
formation in Hardwoods, wood science 6 (1974) 206-211.
Meyer, R. W., Lawrence, L., Shake in coniferous wood: an anatomical study, Forest products
journal 18 (1968) 51-56.
Owen, D. O., Wilcox, W. W., The association between ring shake, wetwood and fir engraver
beetle attack in white fir, Wood and Fiber 14 (1982) 267-280.
Pozzi, F., Studio anatomico comparativo sulle modalità di frattura del legno di castagno (Castanea
sativa Mill.): campioni affetti da cipollatura e campioni rotti meccanicamente, Tesi di Laurea,
Università degli studi di Torino, 1996. pp. 90.
Saya, I., Indagine sulla cipollatura del legno di castagno e di abete bianco, Contributi scientificopratici per una migliore conoscenza e utilizzazione del legno 6 (1962) 37-41.
Shigo, A. L., Ring shake associated with sapsucker injury, US Forest service research paper NE-8,
1963.
Shigo, A. L., Ring and ray shakes associated with wounds in Trees, Holzforschung 26 (1972) 6062.
Tan, D. M., Stanz-Tschegg, S. E.Tschegg, E. K., Models of wood fracture in Mode I and Mode II,
Holz als Roh- und Werkstoff 53 (1995) 159-164.
Section B
[24]
[25]
[26]
57
Thibaut, B., Fournier, M.Jullien, D., Contraintes de croissance, recouvrance différée à l'étuvage et
fissuration des grumes: cas du châtaignier, Forêt Méditerranéenne 16 (1995) 85-91.
Wilkes, J., Anatomy of zones of ring shake in Eucalyptus maculata, IAWA Bulletin n. s. 7 (1986)
3-11.
Wilson, B. F., A survey of the incidence of ring shake in eastern hemlock, Harvard Forest Papers 5
1962.
Section B
59
3
Relationship between ring shake incidence and
earlywood vessel characteristics in chestnut wood
Patrick Fontia, Otto-Ulrich Bräkerb and Fulvio Giudicia
a
b
WSL Sottostazione Sud delle Alpi,
CH-6504 Bellinzona
WSL Swiss Federal Research Institute,
CH-8903 Birmensdorf, Switzerland
IAWA Journal (2002) 23(3): 287-298
Submitted: 31 August 2001
Accepted: 29 January 2002
Published: August 2002
60
Section B
Abstract
This paper investigates whether in chestnut (Castanea sativa Mill.) a relationship exists
between anatomical features of earlywood vessels, which may contribute to weakening
the wood, and the incidence of ring shake. The study compared two groups of 30 wood
discs with and without ring shake, collected in three coppice stands in Southern
Switzerland. Shake-prone stems are not characterised by more numerous and wider
earlywood vessel lumina than the shake-free ones. Hence the hypothesis that ring shake
is favoured by the weakening effect of earlywood cell lumina is rejected.
Castanea sativa / earlywood vessels / wood structure / ring shake / coppice.
Section B
61
Introduction
Chestnut wood (Castanea sativa Mill.) displays excellent technical features: it has good
mechanical properties, high durability and a pleasant appearance. Such qualities make it
one of the most versatile woods available in Europe (Bourgeois 1992). Nevertheless
chestnut wood is today a largely neglected natural and renewable resource. The huge
economical potential of the timber, especially for uses with a high added value, is
strongly obstructed by the widespread presence of ring shake. Ring shake is a circular
failure that develops in the tangential plane of the trunk in the woody tissue between
two contiguous annual rings or within growth rings (Chanson et al. 1989), which
strongly downgrades the timber and prevents further processing.
In the case of chestnut, ring shake failure mainly arises in the earlywood zone, where it
occurs running across the large earlywood vessel lumina which are arranged in rows
along the ring boundary (Figure 1) (Fonti et al. 2002). It is well known that ring porous
wood species, like chestnut, have a reduced radial strength when loaded in the radial
direction (Dinwoodie 1975; Beery 1983). In fact the earlywood zone, which has plenty
of large vessels, represents a natural weak zone and can consequently easily fail. Some
species cope with this weakness thanks to numerous and large wood rays, which
through a high radial tensile strength are able to reinforce the earlywood zone
(Schniewind 1959; Beery 1983; Mattheck et al. 1994; Badel & Perré 1999; Burgert et
al. 1999; Burgert & Eckstein 2001). But this is only partially the case for chestnut
wood, which is characterised by small and narrow, uniseriate wood rays and thus cannot
offer great resistance to failures developing in the tangential plane. The high propensity
of this species to ring shake may therefore be the result of its extreme weakness in the
tangential plane (Ferrand 1980).
Figure 1: Enlarged cross section of ring shake in chestnut wood. The earlywood zone, which has
numerous large vessels lumina arranged in rows along the annual ring boundary, represents a weak zone
if any load is applied in the radial direction and therefore fails brashly (perpendicular to the cell axis)
giving rise to the ring shake.
IAWA Journal (2002), Vol. 23 (3), 287-298
62
Section B
Nevertheless, even with a greater tangential weakness, not every chestnut tree is
affected by ring shake. Analysis of ring shake occurrence within a single coppice stand
has shown that the phenomenon tends to be limited to all stems of a restricted number of
stools (Macchioni & Pividori 1996; Fonti et al. 1998), suggesting us that there is an
individual propensity of the stool to the defect.
On the basis of these observations we were led to presume that, in the same way as
between wood species, variability in anatomical structure within chestnut trees may
affect radial tensile strength properties regulating individual ring shake susceptibility. In
addition, the annual ring structure and ring width are expression of several influencing
factors such as genetics, soil, climate, history of the stand, etc. Therefore the use of
anatomical parameters for discriminating individuals with and without ring shake seems
to be a reasonable approach. Consequently, the objectives of this study are to use
anatomical parameters in order a) to describe the radial variability of some earlywood
vessel lumina features and b) to verify whether ring shake development in chestnut
wood is related to some of these anatomical characteristics. Specifically we hypothesise
that ring shake easily develops in annual rings or, more general, in chestnut trees
characterised by either more numerous, wider, or a higher amount (line density) of
earlywood vessel lumina.
Materials and methods
Plant material
The study material originated from three over-aged chestnut coppice plots located in the
southern part of Switzerland. The choice of the plots was affected by the availability in
the area of simple coppicing. Stand features are described in Table 1. From each stand
discs of about 5 cm in thickness were gathered from the bases (0.5 m above ground) of
at least 50 dominant or co-dominant chestnut stems. Given that genetic effects were not
considered in our investigation, we selected only one stem per stool. After collection the
discs were first air-dried for about one year to a moisture content of 12-15%, then
polished with a 400 grit sand-paper and finally cleaned with a high-pressure water blast
to clean out the wood dust from the earlywood vessels. Ten shake-free discs and ten
discs with extreme ring shake (defined as dried discs displaying more than 50 cm of
total ring shake failure length), with similar disc diameters, were sampled and
compared. Table 2 describes the mean features of the sampled discs. In each stand no
significant differences in mean age and mean ring width were found between discs with
and without ring shake (t-test, p<0.05).
Ring shake incidence
Ring shake incidence was surveyed taking into account all failures noticed on dried
wood discs. In fact, during the drying process further ring shake could develop through
the relieving of internal wood stresses (Chanson et al. 1989). Because this phenomenon
particularly occurs in wood discs that have already displayed ring shake immediately
after tree felling (Cielo 1988; Fonti et al. 1998), the drying process further accentuates
the proneness of the discs to ring shake.
All failures occurring along the annual ring longer than 1 cm were considered as ring
shake. Each single ring shake was then characterised by the length of the failure, the
Relationship ring shake – earlywood vessels
Section B
63
year of the fractured annual ring and the angle of the disc circular sector enclosed by the
ring shake (ring shake circular angle).
Table 1: Stand characteristics
Characteristics
Altitude
Exposure
Slope
Surface
Age
Silvicultural treatment
Stools per ha
Stems per ha
Mean diameter of dominant stems
Mean height of dominant stems
Units
[mASL]
%
Novaggio
710
West
45
Gerra
600
West – South West
50
Bedano
530
South East
25-50
[ha]
0.13
1.57
1.355
[years]
45-50
None
46-60
None
[ha-1]
[ha-1]
[cm]
[m]
470
1263
27.5
17.4
50
Thinning after 32 years
(1986)
235
403
29.9
15.6
447
719
25.6
16.7
Table 2. Descriptive characteristics of the sampled wood discs. Mean age, mean ring width and mean
ring shake incidence of discs among sites and ring shake groups, with and without ring shake (mean ± SE;
1 value for each disc). The disc ring width was calculated as a mean of all its annual ring widths. Mean
age and mean ring width between groups were not significantly different (p<0.05, t-test)
Site
Discs group
Mean age
Mean ring
width
Novaggio
Without ring shake
With ring shake
Without ring shake
With ring shake
Without ring shake
With ring shake
[year]
43.9 ± 1.2
45.8 ± 1.4
44.7 ± 0.2
45.0 ± 0.0
52.0 ± 1.4
52.8 ± 2.1
[mm]
2.251 ± 0.261
2.096 ± 0.170
2.399 ± 0.142
2.926 ± 0.267
3.316 ± 0.303
3.160 ± 0.176
Gerra
Bedano
Total ring
Total ring
Total
shake failure
shake
number of
length
circular
ring shakes
[cm]
on each disc angle [gon]
20.6 ± 3.8
900 ± 137
86.7 ± 14.4
23.6 ± 2.7
1168 ± 145 150.1 ± 22.1
34.5 ± 4.2
1331 ± 136 187.1 ± 22.2
Earlywood vessel features
The number, size and arrangement of the earlywood vessels lumina are supposed to
influence the radial strength capacity of the wood and consequently the susceptibility
against ring shake. Therefore we characterised these features through some parameters
measured on cross section of a small portion of single annual rings. The parameters
considered are:
•
the mean vessel lumen area in mm2 (MVA), which is the average cross-section
lumen area of all the vessel lumina larger than 0.01 mm2 on the considered single
annual ring portion;
•
the total vessel lumen area in mm2 (TVA), which is the sum of the all vessel lumina;
•
the vessel lumen line percent (VL%). This parameter indicates the relative
proportion of vessel lumina along the first row of earlywood vessels. It corresponds
to the ratio between the total length of the earlywood vessel lumina that cross a
tangential line (drawn along the first row of vessels) and the total length of the line
(as shown in Figure 2).
IAWA Journal (2002), Vol. 23 (3), 287-298
64
Section B
Figure 2: Example of a digitised image that has been analysed. Transverse sections were viewed with a
compound microscope using the 8x objective and images were captured through a video camera. Images
were analysed with the “Images Pro Plus” (version 4.0.0.13 for window 95/NT/98) digital analysis
program. Labelled objects have been recognised and ring width measured. The tangential line crossing
the first row of earlywo od vessels is used to measure the line-proportion of vessel lumen (VL%).
Earlywood vessel lumen parameters were measured on the prepared disc cross-section.
Measurements were performed ring by ring along an 8 mm wide radius strip. The radius
chosen for measurements was positioned midway between the longest and the shortest
radii of the disc in order to reduce the undesired effect of stem-eccentricity.
Measurements were automatically performed through an image analysis device
combining a video camera with an 8x magnification and the “Image Pro Plus” digital
analysis program. The data relative to each annual ring were derived from a digitised
image that reflected the structure of a cross-section of wood from the microscope on the
monitor (Figure 2). An essential step for the automatic measurement is the
discrimination of the vessel lumen from the ground tissue by means of different colour
brightness. The image program was set up with filters (morphological 2x2 squares, 1
pass, which erodes the edges of bright objects, and enlarges dark ones) and image
enhancer (equalise, best fit) in order to improve the contrast and to better recognise all
dark objects (vessel lumen). With the aim of avoiding wrong objects identification, by
means of a selecting filter, only dark objects displaying a ratio between horizontal and
vertical axes lower than 2 and with an area larger than 0.01 mm2 were considered.
Smaller vessels are presumed not to influence the ring shake development.
Section B
65
Results and discussion
Ring shake incidence and distribution
Differences between discs with and without ring shake were distinct (Table 2). Ring
shake discs displayed a large number of shakes that on average varied from a minimum
of 20.6 for the wood discs of Novaggio to a maximum of 34.5 for those of Bedano,
where a mean total failure length of 86.7 cm and 187.1 cm, respectively, was observed.
Figure 3: Annual ring shake incidence expressed in total ring shake circular angle by calendar year, in
comparison to the mean radial annual growth. Each graph summarises the data for the 20 discs of each
stand (only 10 of which have ring shake).
IAWA Journal (2002), Vol. 23 (3), 287-298
66
Section B
The data on ring shake distribution of the discs from the three stands (total ring shake
circular angle) were plotted by calendar year, in comparison to the mean radial annual
growth (Figure 3). Failures mainly concentrate in the annual rings belonging to the
central part of the radius confirming previous observations on the radial distribution of
ring shake (Macchioni & Pividori 1996; Fonti et al. 1998). This distribution could be
explained because of increased growth stresses acting in the middle sector of the radius
(Thibaut et al. 1995). Even though along this area some peaks of ring shake were
observed (i.e. in the annual rings 1972-73 for Novaggio’s discs), no clear relationship
between ring shake incidence and the mean annual radial growth was evident. Only for
the discs of Gerra an increase of ring shake was noticed in association to an abrupt
change of radial increment resulting from a silvicultural treatment (thinning). This latter
result is consistent with the reports that more irregularly grown chestnut individuals
produce more ring shakes than more regularly grown ones (Amorini et al. 1998).
Quantitative anatomical parameters of earlywood vessels
The data on earlywood vessel anatomical features relative to each stand (MVA, TVA
and VL%) were plotted to show the radial trend (Figure 4). All the parameters measured
along the radius strips displayed a large range of values. The well-known phenomenon
of juvenile wood (Zobel & van Buijtenen 1989) is clearly expressed only for the MVA,
which increases for about the first 20 years from the pith before levelling off (Figure
4a), but is not plainly evident for TVA and VL% (Figures 4b, 4c). Given this
observation, in order to reduce data variability, further analyses were limited to the
mature wood data (ring number from the pith >20). Considering only the mature wood,
MVA at single discs ranged from a minimum of 0.014 mm2 to a maximum of 0.062
mm2, TVA ranged from 0.37 mm2 to 5.88 mm2 and VL% from 27% to 86%.
Although almost the same pattern was observed in each stand, the three stands revealed
dissimilar values in earlywood vessel features. With the exception of the VL%,
differences between stands were found significant (using an ANOVA model with
“individuals” nested in “stand”, p<0.05, data not shown). This different vessel
anatomical responsiveness between stands may therefore results from effects of
environmental factors on wood development, as already observed for some oak species
(Villar-Salvador et al. 1997).
Comparative analysis among wood discs with and without ring shake
Because of the significant differences between stands noticed in the quantitative
anatomical parameters of earlywood vessels, the comparative analyses between discs
with and without ring shake were performed for each site separately. Results from the
comparative analyses performed are summarised in Table 3. Significant differences in
mean (Wilcoxon signed rank test, p<0.05) were observed for MVA, but only for the
sites of Gerra and Bedano. As Figure 5 shows, along the radial axes the ring shake discs
had continuously smaller MVA than the ones without shake (also in the case of
Novaggio), even if for the first annual rings from the pith (in the juvenile wood)
differences in vessels size are less manifest. The lack of significance of the Novaggio
data could be due to the general low growth rate of the stand. It is therefore possible that
the anatomical response was mainly regulated by the unfavourable growing conditions
masking the expression of the differences in ring shake proneness.
Section B
67
Figure 4a--c: Radial pattern for a. MVA, b. TVA, and c. VL%. Each point represents the mean values
relative to a site (squares Bedano, triangles Gerra, circles Novaggio). Bars indicate the standard error of
mean.
IAWA Journal (2002), Vol. 23 (3), 287-298
68
Section B
Table 3. Results of the comparative analysis among discs. MVA, TVA and VL% of discs distinguishing
among sites and ring shake groups (mean ± SE; 1 value for each disc). Single disc values are calculated as
a mean of the value measured on the mature wood annual rings (ring number from the pith > 20).
Comparison of means was performed using a non-parametrical statistical test (Wilcoxon signed rank test)
Characteristic
MVA [µm2]
TVA [mm2]
VL% [%]
Sites
Novaggio
Gerra
Bedano
Novaggio
Gerra
Bedano
Novaggio
Gerra
Bedano
Discs without ring
shake
(n = 10)
27.52 ± 1.81
38.22 ± 2.06
45.04 ± 0.84
1.373 ± 0.279
1.867 ± 0.118
2.473 ± 0.307
58.4 ± 2.2
49.1 ± 2.1
49.6 ± 1.9
Discs with ring
shake
(n = 10)
26.20 ± 0.92
29.64 ± 1.81
37.21 ± 1.74
1.176 ± 0.114
1.581 ± 0.181
1.984 ± 0.162
58.2 ± 1.7
55.8 ± 1.7
54.0 ± 1.9
p-value
0.676
0.010*
0.001*
0.671
0.096
0.199
0.909
0.023*
0.151
* =P<0.05
It is however surprising to observe that the wood discs with ring shake display
significantly smaller lumen vessel area than the ones without shakes. This contradicts
the initial hypothesis that the development of chestnut ring shake is favoured in trees
characterised by wider earlywood vessels. Instead it suggests there is a relationship
between ring shake and narrow earlywood vessels. However, at this stage of analysis it
is too early to draw the cause-effect relationships between these two elements.
Comparative analysis among annual rings with and without ring shake
belonging to shake-prone discs
Considering that ring shake develops mainly in the central part of the radius (Figure. 3),
comparative analysis among annual rings with and without ring shake was performed
only on rings belonging to this region of the stem. Only data from ring shake discs from
the same stands were taken into account. Of course, the surveys on ring shake annual
rings were not performed in the failure zones but in the surrounding area. In this
analysis the belonging of the annual ring to the individuals was ignored, with the
assumption that each of the single annual rings was an independent observation. Results
shown in Table 4 indicate that there are no differences in mean between the two groups
for all the earlywood features measured. This outcome indicates that among the annual
rings of the same wood disc there is no evidence of distinctive vessel anatomical
features, as for example very small MVA, that seemed to promote the development of
ring shake in the comparison among wood discs.
Section B
69
Figure 5: MVA area against the year of the annual ring for each site distinguishing between discs with
and without ring shake. Closed circles correspond to discs with ring shake and the open circle to the discs
without ring shake. Bars indicate the standard error of mean.
IAWA Journal (2002), Vol. 23 (3), 287-298
70
Section B
Table 4. Results of the comparative analysis among annual rings. MVA, TVA and VL% of annual rings
(mean ± SE) belonging to the central part of the radius (20 < ring number from the pith < 35; ca. 14
values for each disc) of discs with ring shake, distinguishing among sites and annual rings with and
without ring shake. Comparison of means was performed using a non-parametrical statistical test
(Wilcoxon signed rank test).* =P<0.05
Characteristic
MVA [µm2]
TVA [mm2]
VL% [%]
Sites
Novaggio
Gerra
Bedano
Novaggio
Gerra
Bedano
Novaggio
Gerra
Bedano
Annual rings
without ring shake
26.35 ± 0.49
30.01 ± 0.73
36.91 ± 1.07
1.286 ± 0.066
1.253 ± 0.073
2.032 ± 0.118
56.2 ± 0.9
53.3 ± 1.0
53.5 ± 1.0
Annual rings with
ring shake
25.99 ± 0.57
30.11 ± 0.96
36.65 ± 0.94
1.163 ± 0.061
1.486 ± 0.086
2.466 ± 0.143
57.0 ± 1.1
53.3 ± 1.2
52.7 ± 1.0
p-value
0.722
0.946
0.852
0.583
0.020*
0.205
0.638
0.767
0.498
Conclusion
This study provides a valid body of evidence that wood discs with ring shake are not
characterised by wider and more numerous earlywood vessel lumina than the discs
without ring shake. For two of the three sites, MVA was even found significantly
smaller in discs with ring shake, while the TVA and the VL% did not display significant
differences. Therefore the hypothesis that the development of ring shake is favoured by
more numerous and wider earlywood vessel lumina is rejected. While considering the
shake-prone discs only no significant anatomical differences among annual rings with
or without shake were found. Therefore no alternative hypothesis on the effect of wood
structure on the incidence of ring shake in chestnut can be proposed.
Acknowledgements
The technical staff of the Swiss Federal Institute for Forest, Snow and Landscape,
Branch Station South of the Alps participated in collecting the material in the forest.
Nina Lichtfuss and Matteo Buzzi contributed to the assessments with digital images.
Marco Conedera of the Swiss Federal Institute for Forest, Snow and Landscape, Branch
Station South of the Alps reviewed the text. We express our gratitude to them all.
References
Amorini, E., S. Bruschini, M. Fioravanti, N. Macchioni, M.C. Manetti, B. Thibaut & L. Uzielli 1998.
Studi sulle cause di insorgenza della cipollatura nel legno di castagno (Castanea sativa Mill.). In:
Comunità montana delle Prealpi Trevigiane (ed): Atti del convegno nazionale sul castagno,
Cison di Valmarino (Treviso), 23-25 Ottobre 1997: 269--292.
Badel, E. & F.E. Perré 1999. Détermination des propriétés élastiques d'éléments individuels du plan
ligneux du chêne par des essais de traction sur micro-éprouvettes. Annals of Forest Science. 56:
467--478.
Beery, W.H. 1983. Quantitative wood anatomy: relating anatomy to transverse tensile strength. Wood and
Fiber Science. 15: 395--407.
Bourgeois, C. 1992. Le châtaignier: un arbre, un bois. Institut pour le développement forestier. 367.
Burgert, I., A. Bernasconi & D. Eckstein 1999. Evidence for the strength function of rays in living trees.
Holz als Roh- und Werkstoff. 57:397--399
Burgert, I. & D. Eckstein 2001. The tensile strength of isolated wood rays of beech (Fagus sylvatica L.)
and its significance for the biomechanics of living trees. Trees. 15: 168--170.
Chanson, B., J.M. Leban & B. Thibaut 1989. La roulure du châtaignier (Castanea sativa Mill.). Forêt
Méditerranéenne. 11: 15--32.
Section B
71
Cielo, P. 1988. Incidenza e tipologie della cipollatura in un ceduo di Castagno del Comune di Garessio
(CN): ricerche bibliografiche e sperimentali. Tesi di laurea, Facoltà di agraria, Università degli
studi di Torino: 356.
Dinwoodie, J.M. 1975. Timber - A review of the structure-mechanical property relationship. Journal of
Microscopy. 104: 3 --32.
Ferrand, J.C. 1980. La roulure du châtaignier: Note préliminaire. Rapport, INRA - CNRF Station de
recherche sur la qualité du bois: 15.
Fonti, P., F. Giudici, L.J. Kucera, E. Ott & E. Pöhler 1998. Studio sulla cipo llatura in un ceduo castanile.
In: Comunità montana delle Prealpi Trevigiane (ed): Atti del convegno nazionale sul castagno,
Cison di Valmarino (Treviso), 23-25 Ottobre 1997: 293--302.
Fonti, P., N. Macchioni & B. Thibaut 2002. Ring shake in chestnut (Castanea sativa Mill.): state of the
art. Annals of forest sciences. Annals of forest science 59: 129--140.
Macchioni, N. & M. Pividori 1996. Ring shake and structural characteristics of a chestnut (Castanea
sativa Mill.) coppice stand in northern Piedmont (north-west Italy). Annals of Forest Sciences. 53:
31--50.
Mattheck, C., W. Albrecht & F. Dietrich 1994. Die Biomechanik der Holzstrahlen. Allgemeine Forst- und
JagdZeitung. 165: 143--147.
Schniewind, A.P. 1959. Transverse anisotropy of wood: a function of gross anatomic structure. Forest
products journal. 9: 350--359.
Thibaut, B., M. Fournier & D. Jullien 1995. Contraintes de croissance, recouvrance différée à l'étuvage et
fissuration des grumes: cas du châtaignier. Forêt Méditerranéenne. 16: 85--91.
Villar-Salvador, P., P. Castro-Díez, C. Pérez-Rontomé & G. Montserrat-Martí 1997. Stem xylem features
in three Quercus (Fagaceae) species alon a climatic gradient in NE Spain. Trees. 12: 90--96.
Zobel, B.H. & J.P. van Buijtenen 1989. Variation among and within trees. In: Timell T.E. (ed), Wood
variation: its causes and control: 72--131. Springer series in wood science, New York.
IAWA Journal (2002), Vol. 23 (3), 287-298
Section B
73
4
Is ray volume a possible factor influencing ring
shake occurrence in chestnut wood?
Patrick Fontia and Beat Freyb
a
b
WSL Sottostazione Sud delle Alpi, Via Belsoggiorno 22,
Casella postale 57, 6504 Bellinzona, Switzerland
WSL Swiss Federal Research Institute, Zürcherstrasse 111,
8903 Birmensdorf, Switzerland
Trees: structure and function (2002) 16(8): 519-522
Submitted: 12 December 2001
Accepted: 3 May 2002
Published online: 24 July 2002
Published: November 2002
74
Section B
Abstract
Radially oriented ray tissue is important for regulating radial strength of wood. The
present study was undertaken in order to assess whether radial rays influence ring shake
occurrence in chestnut wood (Castanea sativa Mill.), a species very prone to ring shake.
Ray volume fraction was measured on tangential samples from two sets of wood discs,
either with or without ring shake, collected from three coppice stands in the southern
part of the Swiss Alps. Our data indicate that ring shaken trees tend to exhibit higher ray
volume than unshaken ones. This rather unexpected finding could be partly explained if
biomechanical processes that control and determine the inner architecture of the tree are
considered.
Wood rays / Castanea sativa / ring shake / radial strength / biomechanics.
Section B
75
Introduction
Wood rays are known to play a key role in parallel-to-grain cracks occurring in wood.
According to Mattheck and Kubler (1995), wood rays were shown to be structural
points of weakness as well a reinforcement depending on the direction of load: they act
as planes of weakness in wood that is subjected to tensile stress in a tangential direction
but also as a reinforcing structure in wood that is subjected to tensile stress in a radial
direction. The importance of ray tissue for the radial strength of wood has been
demonstrated in several ways. Firstly, mechanical observations have shown that wood
species displaying an elevated proportion of rays exhibit higher values of radial strength
(Burgert et al. 2001; Tschegg et al. 2001; Eckstein et al. 1998; Zipse 1997; Beery et al.
1983; Keller and Thiercelin 1975; Schniewind 1959; Kollmann 1956). Secondly, thanks
to recent technological progress permitting measurement of the radial properties of
single isolated multiseriate wood rays, it has been demonstrated that in beech wood
radial rays are about 10 times stiffer than axial tissues in the radial direction (Badel and
Perré 1999) and about 3 times stronger than entire beech wood in its dry state (Burgert
and Eckstein 2001). Thirdly, observations performed on spruce wood subjected to radial
tensile failure tests and bending tests parallel to the grain have evidenced that cracks
along growth ring borders were frequently arrested or became discontinuous at wood
rays, indicating that breaking a wood ray across its longitudinal axis requires a high
work of fracture (Bodner et al. 1997). Finally, confirming the hypothesis advanced by
Mattheck and Kubler (1995) and by Eckstein et al. (1998), it has also been shown that
living trees are able to adapt to new radial stress by modifying the orientation (Burgert
et al. 1999) and the relative volume (Albrecht et al. 1995) of radial rays. On the basis of
these numerous observations we can therefore conclude that in the radial direction rays
constitute an almost ideal “fibre reinforced” tissue that improves the resistance of wood
against tangential crack initiation and propagation.
Ring shake is a defect in chestnut (Castanea sativa Mill.) that consists of a separation
that mainly forms in the (weakest) tangential plane in the ligneous tissue along the
annual growth ring (Chanson et al. 1989) (Fig.1). Given that reduced radial strength is
thought to be one of the major causes of ring shake (Fonti et al. 2002), it could therefore
be interesting to analyse the relationship between ray characteristics and ring shake
occurrence in chestnut wood. This study explores from an anatomical point of view
whether there are differences in the quantitative ray volume fraction that may help to
explain why some individuals develop the defect while some others, grown under the
same stand conditions, do not.
Material and methods
C. sativa samples were obtained from three mature coppice stands situated in Southern
Switzerland. The last coppicing of the stands was in 1946-1951 for Novaggio, in 1950
for Gerra and in 1940-1955 for Bedano. From each stand discs about 5 cm in thickness
were gathered from the bases (0.5 m aboveground) of at least 50 dominant or codominant chestnut shoots. After collection the discs were firstly air-dried for about 1
year to a moisture content of 12-15%. Afterwards, in order to perform a comparative
analysis, from each site one group consisting of 10 shake-free and another one formed
by 10 extremely ring shaken discs (defined as dried discs displaying ring shake failure
of more than 50 cm total length) were sampled out from all the gathered discs looking
for similar representativeness of disc diameters in both categories. A strip 2 cm wide
Trees, structure and function (2002) , Vol. 16 (8), 519-522
76
Section B
was chosen, including the pith but avoiding knots and other visible defects, across the
minimum diameter of each disc. The minimum diameter was preferred in order to
minimize the presence of reaction wood. After re-wetting wood strips (storing it in
water for 3 days), tangential sections were cut for the yearly ring corresponding to the
growth years 1960, 1967, 1974, 1981, 1988 and 1995 and subsequently refined with a
sliding microtome to obtain a clear surface. Each tangential section was observed with a
low-temperature scanning electron-microscope (LTSEM) (Frey et al. 1996) and five
0.17 mm2 wood-images were randomly captured at a magnification of x 200. In total
1,800 images were collected. In order to measure the relative ray volume fraction a
transparency was applied to each photographed image and rays were manually
transferred to the transparency. Then, after digitisation of the transparency
measurements were performed using the “image pro plus” digital analysis program (Fig.
2), which permitted identification of and measurement of the surface covered by rays
and its relation to the total tangential section surface considered (relative ray volume
fraction).
Figure 1: Example of ring shake developing across the earlywood vessels of chestnut wood. Due to its
particular wood structure, which is characterized by ring-porous wood and by small uniseriate radial rays,
it is reasonable to suppose that the wood is particularly weak against stress acting along the radial axis. In
fact, on one hand the earlywood vessels provide a fragile plane where ring shake can easily develop,
while on the other hand the rays have an important role in offering resistance to cracks developing. This
could therefore explain why, compared to other wood species, chestnut is so inclined to develop ring
shake (Fonti et al. 2002).
Section B
77
The relative ray volume fraction of all chestnut tree samples studied was in the range
between 8.7% and 17.2% with a mean of 12.2% (Tab. 1) confirming the results (Burgert
et al. 2001) that place chestnut wood amongst hardwood species that display a low
volume of rays. Nevertheless these amounts seem to depend on the stand conditions
because differences in mean (p<0.05, ANOVA one-way analysis) were observed
between the three stands.
Figure 2a-c: Measurement procedure to provide images for the analysis: a The tangential section was
captured using LTSEM, b a transparent was overlapped on the captured image and the perimeter of each
single ray was drawn on the transparent. The transparent image was then digitised and used for the image
analysis measurements, c each single ray was recognised, shaded and subsequently measured.
Table 1: Relative ray volume of all trees (ring shaken or not) among sites.
1
The mean value and
standard deviation refer to the mean value of the characteristic for each individual tree.
Relative ray volume (%)
Number of trees
Minimum
Maximum
Mean1
SD1
Gerra
20
8.89
15.28
12.28
1.58
Novaggio
20
8.74
12.23
10.65
1.20
Bedano
20
10.38
17.23
13.59
1.91
All trees
60
8.74
17.23
12.18
1.98
78
Section B
Assuming that radial rays effectively have a strengthening function along the radial axis
of the wood, we therefore expected to observe a higher volume fraction of rays in
chestnut trees without ring shake than in those displaying this defect. However, results
obtained from our study cannot confirm this expectation. As Figure 3 shows, no lower
values between ring shaken and not ring shaken groups within the same site have been
found. In contra, ring shaken trees tended to have a higher volume of rays, even if the
difference in mean was only significant for the stand of Bedano (p<0.05, t-test). This
trend was fully confirmed along the entire radial axis, as shown in Figure 4. Here we
found that the tangential section taken from ring shaken trees always displayed a higher
volume of rays than the unshaken ones, for all three stands.
Gerra
Novaggio
Bedano
Relative rays volume
0.16
0.14
0.12
0.10
0.08
0
1
0
1
0
1
Ring shake group
Figure 3: Box plot of relative ray volume distinguishing between sites and groups: The plot data refers
to the mean value calculated for each tree. 0 = group without ring shake, 1 = group with ring shake.
Comparison of mean (t-test) between group 0 and 1 has given the following P values: Gerra = 0.1118;
Novaggio = 0.3483; Bedano = 0.0074.
Section B
79
19%
GERRA
18%
Ray volume fraction [%]
17%
16%
15%
14%
13%
12%
11%
10%
9%
1960
1967
1974
1981
1988
1995
Tangential section
19%
NOVAGGIO
18%
17%
Wood discs group without ring shake
16%
Wood discs group with ring shake
15%
14%
13%
12%
11%
10%
9%
1960
1967
1974
1981
1988
1995
Tangential section
19%
BEDANO
18%
17%
16%
15%
14%
13%
12%
11%
10%
9%
1960
1967
1974
1981
1988
1995
Tangential section
Figure 4: Mean values and standard deviation of the relative ray volume fraction along the radial axes
(for each tangential section considered) differentiating among stands and wood discs group with and
without ring shake.
80
Section B
In attempting to interpret this surprising result, we have to consider the biomechanical
processes stimulating the ring shaken trees to produce more rays than the unshaken
trees. In this case it is plausible that, in order to face higher stress levels and therefore
confront the problem of ring shake, growth-stress-overloaded trees might generally
create more rays than the stress-poor ones. But when the tree falls and the internal wood
balance is destabilized, some of the stress can be released (Chanson et al. 1989) and
new cracks can develop. In fact, most ring shakes have already occurred at that time, i.e.
in fresh fallen timber (Fonti and Macchioni, unpublished data). The stress-overloaded
trees (those with a higher ray volume fraction) then have more stress to release, which
in some cases is so strong that the enlarged volume of rays may not be able to
counteract. The expected outcome is therefore consistent with that observed in this
study, even if apparently contradictory, where trees with ring shake display a larger
volume fraction of radial rays than the unshaken ones.
Our study showed that the ray volume fraction does not directly influence the
development of ring shake in chestnut wood. However, the results suggest that the ray
volume fraction could likely be an indicator of the amount of growth stress which
occurred in the living trees, and which could play an important role in the development
of ring shake in chestnut.
Acknowledgements
We thank M. Conedera and F. Giudici of the WSL Sottostazione Sud delle Alpi for their
valuable advises and constant help given during the work.
References
Albrecht W, Bethge K, Mattheck C (1995) Is lateral strength in trees controlled by lateral mechanical
stress? J Arboric 21:83-87
Badel E, Perré P (1999) Détermination des propriétés élastiques d'éléments individuels du plan ligneux du
chêne par des essais de traction sur micro-éprouvettes. Ann For Sci 56:467-478
Beery H, Geza I, Mc Lain E (1983) Quantitative wood anatomy - relating anatomy to transverse tensile
strength. Wood Fiber Sci 15: 395-407
Bodner J, Schlag M-G, Grüll G (1997) Fracture initiation and progress in wood specimens stressed in
tension. I. Clear wood specimens stresses parallel to the grain. Holzforschung 51:479-484
Burgert I, Eckstein D (2001) The tensile strength of isolated wood rays of beech (Fagus sylvatica L.) and
its significance for the biomechanics of living trees. Trees 15:168-170
Burgert I, Bernasconi A, Eckstein D (1999) Evidence for the strength function of rays in living trees.
Holz Roh- Werkstoff 57:397-399
Burgert I, Bernasconi A, Niklas K, Eckstein D (2001) The influence of rays on the transverse elastic
anisotropy in green wood of deciduous trees. Holzforschung 55:449-454
Chanson B, Leban J-M, Thibaut B (1989) La roulure du châtaignier (Castanea sativa Mill.). For Mediterr
11:15-32
Eckstein D, Burgert I, Schwab E (1998) Gibt es einen Zusammenhang zwischen der radialen und der
axialen Festigkeit im lebenden Baum? Allg Forst Jagdztg 169:101-103
Fonti P, Macchioni N, Thibaut B (2002) Ring shake in chestnut (Castanea sativa Mill.): state of the art.
Ann For Sci 59:129-140
Frey B, Scheidegger C, Günthardt-Goerg MS, Matyssek R (1996) The effects of ozone and nutrient
supply on stomatal response in birch leaves (Betula pendula) as determined by digital image
analysis and X-ray microanalysis. New Phytol 132:135-143
Section B
81
Keller R, Thiercelin F (1975) Influence des gros rayons ligneux sur quelques propriétés du bois de hêtre.
Ann For Sci 32:113-129
Kollmann F (1956) Untersuchungen über die Querzugfestigkeit der Hölzer. Forstwisse Centralbl 75:304318
Mattheck C, Kubler H (1995) Wood - the internal optimization of trees. Springer, Berlin Heidelberg New
York
Schniewind A (1959) Transverse anisotropy of wood: a function of gross anatomic structure. For Prod J
9:350-359
Tschegg EK, Frühmann K, Stanzl-Tschegg SE (2001) Damage and fracture mechanism during mode I
and III loading of wood. Holzforschung 55: 525-533
Zipse A (1997) Untersuchungen zur lastgesteuerten Festigkeitsverteilung in Bäumen. Wissenschaftliche
Berichte FZKA 5878, Fakultät für Maschinenbau der Universität Karlsruhe
Section B
83
5
Radial split resistance of chestnut earlywood and
its relation to the ring width
Patrick Fontia and Jürgen Sellb
a
b
EMPA, Swiss Federal Laboratories for Materials Testing and Research,
Wood Department, Ueberlandstrasse 129, CH-8600 Duebendorf, Switzerland
Wood and Fiber Science (in press)
Submitted: 26 February 2002
Accepted: 24 September 2002
84
Section B
Abstract
New equipment was developed to measure the maximal radial split resistance of
individual annual rings in green European chestnut wood (Castanea sativa Mill.). This
equipment was then used to compare the split resistance in chestnut trees with and
without ring shake taken from three differently managed coppices from the southern
part of the Swiss Alps. Results indicate that within these stands radial split resistance
and annual ring width are positively correlated, and that the rates of ring shake
occurrence increase with narrow and weak growth rings. Forest management of chestnut
coppices that leads to an increase in growth thickness might, therefore, be a way of
reducing the risk of ring shake.
Castanea sativa / European chestnut / radial split resistance / ring shake
Section B
85
Introduction
Chestnut wood (Castanea sativa Mill.) is commonly affected by ring shake, a serious
problem reducing the use of this wood for high-added-value products (Bourgeois 1992).
Ring shake is a type of wood crack which develops in the tangential plane of the trunk
growing parallel to the annual ring (Chanson et al. 1989). Due to the high incidence of
ring shake and difficulties in processing, the demand for this wood is limited.
Chestnut wood research has been principally aimed at avoiding the damage caused by
ring shake, either by improving wood processing and uses or by identifying the factors
controlling the development of failures. In particular research focused on the causes
leading to the more problematic “healthy” type (as defined by Chanson et al. (1989)),
whose fracture appears to be unrelated to any anomalies of the wood tissue
(discoloration or decay due to trauma, fungal or bacterial activity). Fonti et al. (2002b)
recently reviewed the state of the art on chestnut ring shake. Of particular interest is
Ferrand’s theory (Ferrand 1980), advancing the hypothesis that chestnut wood is prone
to developing ring shake because of its structure that forces the propagation of
tangential rather than radial cracks. It is proposed that the earlywood structure with its
ring porosity and large vessels generates a plane of weakness along the annual ring,
while the thin uniseriate wood rays have a limited ability to resist crack openings in the
tangential plane (Tschegg et al. 2001; Burgert et al. 1999; Eckstein et al. 1998; Beery et
al. 1983; Keller and Thiercelin 1975; Schniewind 1959; Kollmann 1956). In addition,
radially growing cracks, which usually form along the radial rays, have to cross both the
weak earlywood and the tough latewood, which offers resistance to the fracture
propagation without finding preferential paths (along large radial rays). Despite this
high propensity of the species, not every tree is affected by ring shake. Intra-specific
variability in the wood structure may affect radial tensile strength properties regulating
individual (or even annual ring) susceptibility. Even if earlywood specific gravity and
growth rate of ring porous hardwood does not seem to change with ring width (Zobel
and Van Buijtenen 1989), experimental data on earlywood radial strength in relationship
with the growth rate are still inadequate.
Many earlier studies on radial strength have revealed that trunks affected by ring shake
usually display lower average strength values than those without ring shake (Macchioni
1995; Frascaria et al. 1992; Leban 1985) although, because of the wide variability
observed, it was not possible to find a significant relationship between radial weakness
and ring shake. Nevertheless, these earlier studies suffer from some limitations in the
testing methods because they do not allow either comparable data to be collected or the
earlywood radial cohesion strength of wood to be properly characterized. Tensile radial
tests performed by Leban (1985) only allowed one measurement per radial specimen,
without it being known exactly where the specimen would break. Bending tests carried
out by Frascaria et al. (1992) permitted several measurements along the same radial
specimen, but it was not possible to check the exact crack location in this case.
Macchioni (1995) developed a testing method based on torsion load on wood cores.
This method permitted multiple measurements along the same cores and enabled the
load location to be checked. However, the fracture displayed features differing from
ring shake, not occurring precisely in the tangential plane. The wedge splitting
technique patented by Tschegg (1986) permits the radial split resistance to be
determined while avoiding previous limitations. However, the shape and dimensions of
Wood and fiber science, in press
86
Section B
the test specimens are unsuitable for easy measurement of the split resistance of annual
rings positioned in close proximity.
In the present study, an effort was made a) to develop a split testing procedure to obtain
comparable earlywood split resistance measurements for as many annual rings
belonging to the same radial wood specimen as possible. The hypothesis is that the split
resistance is a good indicator of the radial cohesion strength of wood within single
growth rings. In addition, a preliminary study was performed b) to verify whether there
is a relationship between radial wood strength, ring width and ring shake.
Wood materials
Overstory trees were selected in three chestnut coppice stands managed with different
intensity and located in the Swiss canton of Ticino south of the Swiss Alps: 5 trees from
a regularly and intensively thinned stand (Gorduno, 29 year old stands), 21 trees from
an occasionally and partially managed stand (Brione, 26-45 y.o.), and 24 trees from an
unmanaged and abandoned stand (Bedigliora, 62 y.o.). From each of the selected trees,
a 30 cm thick stem disk was cut at 1.3 m above ground stem height. Since green wood
was needed for the splitting test, these disks were immediately taken to the preparation
laboratory. There, every thick wood disk was further sawn in order to obtain 3 to 5
green wood disks (3 cm thick). One of these was used for ring shake observations, one
for the preparation of the wood specimens for the mechanical test, and the remaining
disks were either stored as reserve or used for repeated measurements in order to verify
the consistency of the method used. From the disks for the mechanical testing, two pithto-bark radial strips (2cm wide x 2.5cm high x disk radius length) that displayed the
annual rings well disposed perpendicular to the radius axis were sawn for measurement
of the annual ring width. Since the mechanical tests were performed later in a second
test stage, all these still green wood strips were wrapped in aluminum paper and stored
in a freezer at -17°C in order to prevent desiccation. Prefreezing should not have
affected results; as observed by Stanzl-Tschegg et al. (1994) the fracture toughness
between spruce wood with and without prefreezing did not change significantly.
The splitting test
Device and procedure settings
The splitting test device consists of a universal testing machine (ZWICK type 1474)
provided by a wedge with an angle of 60° and a side length of 2.5cm and employed in
its compression mode (Figure 1). The load capacity of the machine is 100 kN and the
capacity of the load cell used was 100 kN. The loading speed was set up to 50 N/s for
all tests.
Radial split resistance and ring width in chestnut
Section B
87
Splitting wedge
Fmax
Earlywood area
L
F max
T
60°
7 mm
2.5 cm
2 cm
R
3 mm
Starter-Notch
Figure 1: Specimen details and test facilities used for the splitting test.
Set up of wood specimen:
The previously collected pith-to-bark radial strips were set up in order to perform a
maximal number of tests on the same radial strip. Since a minimal distance between two
consecutive splitting tests is needed in order to avoid wood structural damage, we
arbitrarily decided to perform the measurements, when possible, on every fourth annual
ring. For every selected annual ring a 7 mm deep and 3 mm thick starter notch was cut
in the earlywood zone using a special milling head, alternatively changing from the top
to the bottom of the wood-strip specimen (Figure 1). Annual rings displaying either
(partial) failures or an excessive curvature, i.e. these rings where the starter notch did
not entirely cover the earlywood area, have not been considered. Finally, in order to
reduce inelastic compression on the upper edge of the notch during load transmission,
the upper side of the starter notch was slightly rasped with a triangular lime (60° angle)
to increment the contact surface between wedge and wood. Before running the test
wood strip specimens were stored at room temperature for a few hours for thawing.
Splitting tests were performed only on water-saturated wood samples.
Estimation of the splitting resistance
Using the TestXpert software (Version 6.01, produced by Zwick GmbH & Co.
Germany) connected with the device, we recorded the wedge load-displacement curves.
Then the splitting resistance (Rsplit ) was calculated according to Stanzl-Tschegg et al.
(1995) as:
Rsplit = (F max*0.5tgα)/A [N/mm2]
(1)
where Fmax is the maximal load of rupture, α the wedge angle and A the fractured
surface.
Ring shake observation
Ring shake incidence was surveyed taking into account all failures (healthy ring shake)
longer than 1 cm noted on the green wood disks. The length of the failure, its location
on the wood disk and the year of the fractured annual ring were recorded for each ring
shake. Then a ring shake index (Icip) based on ring shake failure length was calculated
for the whole wood disk.
88
Section B
Results
Growth characteristics and ring shake incidence of the disks
As expected, all trees from Bedigliora show slow growth (average disk ring width
<4mm) in contrast to those from Gorduno, which have grown faster, whereas in Brione
both classes are present: 17 slow-growing trees and 6 fast-growing ones (Table 1).
Among the disks selected, we observed that 19 (38%) are shake-free (Icip < 30 cm), 16
(32%) are slightly ring shaken (Icip 30-150 cm) while 15 (30%) are strongly ring
shaken (Icip > 150 cm). Ring shake was mainly located in the middle third of the radius.
It is interesting to note that shake-free disks are present in all growth categories whereas
fast-growing trees are only rarely ring-shaken.
Table 1: Disk classification according to stand, growth rate and shake-incidence. SG = slow-growing,
i.e. average disk ring width < 4mm; FG = fast-growing, i.e. average disk ring width > 4mm, Icip = ring
shake index. Shake incidence groups do not statistically differ in disk diameter (ANOVA, p=0.98)
Stand / N° of
discs
Bedigliora
Brione
Brione
Gorduno
Total
Growth rate
Shake-free
(Icip = 0-30cm)
SG
SG
FG
FG
8
2
4
5
19
Slightly shaken Strongly shaken
(Icip 30–
(Icip>150cm)
150cm)
8
8
8
5
2
16
15
Radial split resistance
Radial split resistance tests were performed in the earlywood area of 1076 annual rings
from 50 trees collected in 3 differently managed coppice stands. Radial split resistance
values generally exhibited wide variability, ranging from a minimum of 0.06 N/mm2 to
a maximum of 2.55 N/mm2 with an average value of 0.86 N/mm2 (q25 =0.55 N/mm2;
q75=1.10 N/mm2).
In order to validate the experimental method, split-test measurements were replicated in
the same annual rings of three adjacent pith-to-bark radial wood specimens from 8 trees.
Altogether 110 replicas were tested. The average standard deviation within replicas was
0.11 N/mm2. Compared to the variability between annual rings belonging to the same
radial strip (average standard deviation =0.32), the variability within the replicas was
significantly smaller (one sided, paired t-test, p < 0.001). This result suggests that the
splitting method used is consistent and sensitive enough for such kind of measurements.
Relationship between radial split resistance, ring width and ring shake
incidence
Figure 2 summarizes radial split resistance, ring width and ring shake incidence
measurements performed in a single tree (Brione no. 3). On this disk it is possible to
qualitatively observe that:
• wider annual rings show higher values of radial split resistance;
Section B
89
• ring shake mainly occurs in the central part of the radius and is more frequent in
sector B, which is characterized by rings displaying narrower and lower radial
cohesion strength than those of sector A.
Similar results are even more apparent when considering all the samples. As illustrated
in Figure 3, which shows the relationship between ring width and radial strength, the
splitting resistance within the 3 stands is significantly related to the width of the annual
ring. The larger the annual ring, the more split-resistant the earlywood is in its tangential
plane. In line with a 1 mm radial growth increment, the split resistance increases from
0.08 N/mm2 for the Gorduno disks to 0.16 N/mm2 for the samples from Brione. In
Figure 3 it is also possible to observe that in the majority of the cases the strength values
of the annual rings located in the proximity of ring shake failure are among the weakest
values measured in rings with similar width. However this does not mean that all
growth rings which showed lower strength values are in the proximity of ring shaken
areas, also because there are some areas where ring shake only rarely occurs (near the
pith and close to the bark). It has also been observed that within the same disk ring
shake is predominantly located in the sector with narrower rings. From a selection of
the disks where the mean ring width rate between the 2 selected disk sectors was clearly
detectable (i.e. mean ring width differing for more than 0.5 mm), it was observed that in
14 out of 18 cases the narrower and also weaker radius showed a higher incidence of
ring shake. According to the binomial distribution β(18, 0.5) with a significance level of
0.05 this rate has a probability p significantly different from 0.5. An analogous trend
was also detectable when considering wood disks with different ring shake severity.
Figure 4 shows the relationship between radial split resistance and annual ring width
calculated for each wood disk. Among the faster-growing trees (average ring width > 4
mm) there are no ring-shaken disks, whereas with slow-growing trees the frequency of
ring-shaken disks increases.
90
Section B
North
Radius A in
disk sector A
55 65 75
85 95
Brione 3
West
Radius B in
disk sector B
East
Ring shake
South
2.5
Ring Width [mm/100]
900
Radius A
Radius B
Fmax A
Fmax B
Ring shake A
Ring shake B
800
700
600
2
1.5
500
400
1
300
200
Rsplit [N/mm2]
1000
0.5
100
2000
1995
1990
1985
1980
1975
1970
1965
1960
0
1955
0
Figure 2: Example of measurement performed on a single wood disk. Above a schematic disk
representation with ring shake and wood specimen location identified. Below the graph shows ring width,
split resistance and ring shake presence of the two wood disc sectors of the specimens considered.
Section B
91
Bedigliora
2.5
Rsplit [N/mm2]
2
1.5
y = 0.12x + 0.38
2
R = 0.30
1
0.5
0
0
2
4
6
8
Ring Width [mm]
10
12
Brione
2.5
Rsplit [N/mm2]
2
1.5
y = 0.16x + 0.52
2
R = 0.44
1
0.5
0
0
2
4
6
8
Ring Width [mm]
10
12
Gorduno
2.5
y = 0.08x + 0.50
2
R = 0.45
Rsplit [N/mm2]
2
1.5
1
0.5
0
0
2
4
6
8
Ring Width [mm]
10
12
Figure 3: Relationship between radial split resistance and ring width according to stand origin. plus =
slow-growing tree values; stars = fast-growing tree values; squares = split measurements performed in
proximity of ring shake. If in the disk sector ring shake occurred in the annual ring preceding and
following the annual ring to be measured, then this measurement has been classified as “in proximity of
ring shake”. Regression lines do not consider split measurement performed in proximity of ring shake.
The slope of the regressions are high significantly different from zero (p-value < 0.001).
92
Section B
7
Shake free
Slightly shaken
6
Strongly shaken
2
Rsplit [N/mm ]
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
Ring width [mm]
Figure 4: Relationship of the radial split resistance and ring width. R2= 0.53. Each point corresponds to
mean values measured for each wood disk. Bars indicate the standard deviation of values. Disks have
been classified in three ring shake incidence groups: circles = shake free (Icip 0-30 cm); triangles =
slightly shaken (Icip 30-150 cm); squares = strongly shaken (Icip > 150 cm).
Discussion and conclusion
The splitting method used may not be ideally suited to an absolute and accurate
characterization of the wood split resistance. More sophisticated testing devices, such as
that patented by Tschegg (1986), are more effective in reducing wedge friction during
load transmission. As the purpose of the study was to compare the radial cohesion of
different annual rings, the device used was sufficiently sensible and consistent, as
demonstrated with the 110 triple replication tests performed on 8 different trees.
The results provided clear evidence of a positive correlation between the annual ring
width and radial wood strength. This trend was supported by the comparison between
wood disks, by the pith-to-bark wood specimens belonging to the same disk, and also
by the comparison of single growth rings. The fact that wider rings seem to be less
prone to shake might suggest that growth rate affects earlywood (anatomical)
characteristics. Nevertheless earlywood width of ring porous species does not change
with ring width (Zobel and van Buijtenen 1989). Whereas considering earlywood
vessels and radial rays features it has been observed that, although an increase in vessel
lumina and ray volume with mean ring width of the stand has been observed, the
differences between ring shaken and shake free disks are not as we initially expected: in
fact ring shaken disks displayed significantly smaller earlywood vessel lumina (Fonti et
al. 2002a) and a higher amount of rays (Fonti and Frey 2002) than the unshaken ones.
However, ring shake development was directly related to weak radial cohesion and
narrow annual growth rings. Ring shake can also develop in wood with high radial
cohesion as well as in large growth rings, but the frequency of shakes tends to be lower
in those rings. In particular, we observed that ring shake very rarely occurred in rings
wider than 4 mm. This does not, however, mean that all narrow rings or slow-growing
Section B
93
trees are always affected by ring shake. In addition, near the pith and close to the bark
ring shake only rarely occurs as observed both in this and in previous studies (Fonti et
al., 2002b). In fact, most narrow rings and about 25 percent of all slow-growing trees
considered for this study are shake-free. As well as radial strength, growth stresses in
wood are also very important with respect to the development of ring shake (Fonti et al.,
2002b). It may, therefore, happen that ring shake occurs in wood with a high radial
strength and vice-versa.
An increase in radial growth corresponds to an improvement in the radial split
resistance, which plays an important role in the development of ring shake.
Enhancement of the radial wood cohesion strength should therefore help to reduce the
risk of the occurrence of ring shake.
Acknowledgements
The authors thank Michele Wildhaber and Athos Maestrini for helping in stand
recognition. We also thank Franco Fibbioli and Daniel Heer for valuable assistance in
wood collection and specimen preparation. Special thanks are due to Kurt Weiss, who
assisted in designing the testing instrument. We also acknowledge Marco Conedera and
Fulvio Giudici for their critical reading of the manuscript and the valuable advice given.
Thanks are also addressed to the two reviewers for their comments on earlier versions of
the manuscript.
References
Beery. H, , Ifju,G. and Mc Lain, E (1983). “Quantitative wood anatomy - relating anatomy to transverse tensile strength.” Wood and
Fiber Science 15(4): 395-407.
Bourgeois C (1992). Le châtaignier: un arbre, un bois, Institut pour le développement forestier.
Burgert I, Bernasconi, A and Eckstein, D (1999). “Evidence for the strength function of rays in living trees.” Holz als Roh- und
Werkstoff 57: 397-399.
Chanson B, Leban, J.-M.and Thibaut, B (1989). “La roulure du châtaignier (Castanea sativa Mill.).” For Med. 11: 15-32.
Eckstein D, Burgert, I and Schwab, E (1998). “Gibt es einen Zusammenhang zwischen der radialen und der axialen Festigkeit im
lebenden Baum?” Allgemeine Forst und Jagdt Zeitung 169(6-7): 101-103.
Ferrand J. (1980). La roulure du châtaignier: Note préliminaire, INRA - CNRF Station de recherche sur la qualité du bois: 15.
Fonti P, and Frey, B (2002). “Is the ray volume a possible factor influencing ring shake occurrence in chestnut wood?” Trees structure and function 16(8): 519-522.
Fonti P, Bräker, O-U and Giudici, F (2002a). “Relationship between ring shake incidence and earlywood vessels characteristics in
chestnut wood.” IAWA 23(3): 287-298.
Fonti P, Macchioni, N and Thibaut, B (2002b). “Ring shake in chestnut (Castanea sativa Mill.): state of the art.” Annals of Forest
Science 59(2): 129-140.
Frascaria N, Chanson, B, Thibaut, B and Lefranc, M (1992). “Génotypes et résistance mécanique radiale du bois de châtaignier
(Castanea sativa Mill.): Analyse d'un des facteurs explicatifs de la roulure.” Annals of Forest Science 49: 49-62.
Keller R and Thiercelin, F (1975). “Influence des gros rayons ligneux sur quelques propriétés du bois de hêtre.” Annals of Forest
Science 32(2): 113-129.
Kollmann F (1956). “Untersuchungen über die Querzugfestigkeit der Holzer.” Forstwissenschaftliches Centralblatt 75: 304-318.
Leban J. (1985). Contribution à l'étude de la roulure du châtaignier. Nancy, Institut National Polytechnique de Lorraine: 164.
Macchioni N (1995). “Mechanical strength and ring shake in chestnut (Castanea sativa Mill.).” Forêt Méditerranéenne 16: 67-73.
Schniewind A.(1959). “Transverse Anisotropy of wood: a function of gross anatomic structure.” Forest Products Journal 9: 350-359.
Stanzl-Tschegg, S., Tan, D. and Tschegg, E. (1995). "New splitting method for wood fracture characterization." Wood science and
Technology 29: 31-50.
Stanzl-Tschegg, S., Tschegg, E. and Teischinger, A. (1994). "Fracture energy of spruce wood after different drying procedures."
Wood and Fiber Science 26: 467-478.
Tschegg E.(1986). "Equipment and appropriate specimen shape for tests to measure fracture values." Patent n° AT-390328
Tschegg, E., Frühmann, K. and Stanzl-Tschegg, S. (2001). “Damage and fracture mechanisms during mode I and III loading of
wood.” Holzforschung 55 (5): 525-533.
Zobel, B.H. and van Buijtenen, J.P. (1989). Wood variation: its causes and control. Springer series in wood science, Springer, New
York, 363 pp.
Section B
95
6
Growth strain and ring shake in chestnut coppice
trees
Patrick Fontia
a
Forêt méditerranéenne (submitted 2002)
Submitted: 4 July 2002
96
Section B
Abstract
Mechanical failures in wood are the result of an imbalance between wood strength and
stresses. European chestnut is a wood species very often affected by ring shake, a
concentric failure parallel to the annual growth ring. The timing and the stem position in
which ring shakes occur let us suppose that, combined with a radial structural weakness,
growth stresses result in ring shake development. However, to this date no evidence of a
relationship between high stress level and elevated ring shake incidence in chestnut
wood has been supported. To investigate growth stresses, longitudinal displacements
induced by stress release were estimated at the periphery of the stem using the single
hole drilling method. To examine ring shake, the length and position of splits occurring
on cut sections near stress measurements (breast height) were surveyed. Results do not
show any clear relationship between the level of strain-induced displacement (SID) and
ring shake intensity. However, due to some limitations of the applied method in
estimating stress levels, it is not possible to asses or reject whether ring shake incidence
is related to elevated growth stress.
Growth stresses / Castanea sativa / ring shake / single hole drilling method
Section B
97
Introduction
The wood in standing trees undergoes internal stress during the entire life of the tree.
This stress, commonly named “growth stress”, originates in maturation strains and is
impeded by the mass of the entire trunk (Kubler, 1987). When growth stresses are
relieved by felling and sawing, the wood may split or twist and lose value (Figure 1)
(Archer, 1986).
European chestnut (Castanea sativa Mill.) is very often affected by ring shake, a split
parallel of the annual ring that develops in the longitudinal-tangential plane of wood. If
on one hand chestnut wood structure is prone to this kind of fracture, then on the other
hand there must be a driving force (stresses) that triggers the development of this defect.
Observations performed on the timing in which ring shake appears as well as their
position and the anatomical characteristics they often display (Fonti and Macchioni,
submitted) let us suppose that the release of growth stresses plays an important role in
the mechanism leading to ring shake.
The aim of this study is to verify if there is a relationship between growth stresses and
the development of ring shake in chestnut wood. Because direct measurements of
stresses in wood are not possible, the longitudinal strain measurement on the tree’s
surface will be used as a stress indicator. The longitudinal displacements have been
related to the incidence of ring shake occurring on the stem cross-cut wood disc.
Periphery
Crosscut
Pith
+
+
-
-
Figure 1: Example of stress relieving due to crosscutting. Above: Exaggerated dimensional changes in
a log as a consequence of the release of longitudinal growth stresses. The core expands while the
periphery contracts. Below: Distribution of longitudinal growth strain in stems. Tension areas marked +;
compression areas -.
Forêt méditerranéenne, submitted 2002
98
Section B
Selected trees and wood discs
The performed study is mainly based on the analysis between SID measurements on
standing trees and the ring shake incidence occurring on the disc cross-section gathered
in the proximity of the accomplished SID survey. The chosen trees were selected among
overstory trees grown in three chestnut coppice stands located in the Swiss Canton of
Ticino south of the Swiss Alps: 5 trees from a regularly and intensively thinned stand
(Gorduno, 29 year old stands), 21 trees from an occasionally and partially managed
stand (Brione, 26-45 years.old), and 24 trees from an unmanaged and abandoned stand
(Bedigliora, 62 years old).
On each of the selected trees, SID measurements were performed on the standing trees
and then, after the tree’s felling, a 5 cm thick stem disc was cut at 1.3 m above ground
stem height, in the proximity of the SID measurements. Collected wood discs were then
used for the survey of tree-ring growth chronologies and for the characterization of ring
shake incidence.
SID-measurement
The method used for the measurement of the longitudinal SIDs on the wood surface of
standing trees was developed by Fournier et al. (1994). This method is based on the
measurements of dimensional changes occurring in fibre direction as a consequence of
stress release when a hole is drilled on the stem surface (Figure 2).
Figure 2: View of the gauge before and immediately after the drilled hole. As a consequence of the
drilled hole, stresses can be partially relieved. The distance between the two nails has increased of 55 µm,
as indicated by the gauge.
Measurements were performed on each selected tree at breast high on 8 equidistant
points marked around the debarked tree stem’s periphery. The gauge used for SID
measurements was positioned on two 45 mm distant nails, one at the top and one at the
bottom of the point of measurements, and then the distance between pins was re-set to
zero. Next, with the aid of a drill, a cavity 20 mm in diameter was made between the
Growth strain and ring shake incidence
Section B
99
two nails. The hole was considered to be deep enough when the value indicated by the
gauge was stable (about 15 to 25 mm deep). And finally, once the drill was removed,
the longitudinal SID indicated by the gauge was recorded.
On-disc observations
In order to obtain a smooth surface on the disc cross-section the wood was polished
with a 150 grit sand paper. Tree-ring growth chronologies were carried out on each disc
(on two radii) using an apposite digital positioning table for tree-ring analysis (Lintab)
connected to a tree-ring data-managing software (Tsap).
Ring shake incidence was surveyed twice taking into account all “crack-form failures”
(as defined by Fonti and Macchioni (submitted) and Fonti et al. (2002)) longer than 1
cm visible on the wood disc cross-section: once on the green discs and then repeated on
the same wood discs that had been air dried under shelter for about 1 year. The length of
the failure, the location on the wood disc and the year of the fractured annual ring were
recorded for each ring shake. Then a ring shake index (total ring shake circular angle),
which corresponds to the sum of the angles of the disc circular sectors enclosed by ring
shakes, was calculated for the whole wood disc.
SID per stand
Table 1 shows average and standard deviation of the SID levels observed in the single
stands. The observed mean values, which range from a minimum of 37.5 µm to a
maximum of 49.0 µm, are not higher than those measured in other shaking species like
beech, eucalyptus and poplar (Fournier et al., 1994; FAIR-Project 2001). Compared to
other surveys performed on chestnut coppices, measured SID values are lower than
those observed by Thibaut et al. (1995), which on the average were 76 µm for the stand
of Poitou-Charente (France) and 66 µm for Torre Canavese (Piedmont, Italy).
Table 1: SID’s mean, maximal, minimal and standard deviation of the chestnut trees. Data are grouped
per stands and expressed in µm.
Stand
Gorduno
Bedigliora
Brione
Nr. of
trees
5
24
19
Mean
Max
Min
47.6
37.5
49.0
71.3
71.0
91.1
25.5
12.3
24.6
Standard
deviation
12.8
9.8
27.4
The stem’s peripheral distribution of longitudinal displacement for the different stands
is shown in Figure 3. The shape of the curves is always sinusoidal with a more or less
clear peak. The reason for this could be the intensity and main direction of the slope that
may have an effect on leaning of the stem and on crown eccentricity. In fact we observe
a lower SID value in correspondence to the main direction of the slope.
100
Section B
100
90
80
70
SID [ µm]
60
50
40
30
20
10
0
North
N-E
East
S-E
South
S-W
West
N-W
Cardinal orientation of the measurement performed on stem's circunference
Gorduno
Bedigliora
Brione
Figure 3: Distribution of the average surface longitudinal SID for the different stands. Arrows indicate
the main direction and intensity (arrow-length) of stand’s slope.
100
90
Mean tree strain [mm]
80
70
60
y = 6.8905x + 27.353
2
R = 0.3673
50
40
30
20
10
0
0
1
2
3
4
5
6
7
Mean ring width for the last 10 annual growth ring [mm]
Figure 4: Relationship between mean tree SID and mean ring width of the last 10 annual rings.
8
Section B
101
Intra- and inter-tree variation of SID
As shown in Table 1 the mean SID value of trees belonging to the same stand vary
strongly. In the stand of Bedigliora for example, the ratio between the higher (71.0 µm)
and lower (12.3 µm) main tree SID level is 5.7. Intra-tree SID variations are also
conspicuous: some trees displayed stem sectors with longitudinal deformation 5 times
greater than the “standard values”. These observations are similar to those performed in
previous studies (Thibaut, 1994; Elzière, 1995; Thibaut et al., 1995). This phenomenon
is likely related to the heterogeneous distribution of reaction wood, which compared to
“normal” wood possesses a particularly high maturation stress (Fournier et al., 1994).
Relationship between SID and ring width
The analysis relating longitudinal deformation value with the radial growth of the last
annual ring has evidenced that trees displaying larger ring width have higher
deformation value than those with smaller rings (Figure 4). However, contrary to results
from Thibaut (1994), the same trend was not observed within the same trees: in fact
stem sectors characterized by small annual rings do not show clear evidence of lower
longitudinal surface deformation values than sectors displaying large annual rings.
Relationship between SID and ring shake incidence
As shown in Figure 5, no clear relationships between high tensile strain in the outer part
of the stem and ring shake incidence observed on both the green and dried wood discs
was found: the extension and frequency of ring shake is not clearly related to the
maximum value of displacement at stress release, as shown in Figure 6. In fact, ring
shake can be largely and homogeneously distributed all around the disc in trees that
displayed low SID levels (6b) and can be nonexistent in trees with high SID levels (6a).
It was also observed that in heterogeneously distributed SID trees ring shake can occur
in the sectors displaying either high (6c) or low SID level (6d).
500
Total ring shake circular angle [gon]
450
400
350
300
250
200
150
100
50
0
0
10
20
30
40
50
60
Mean stress [mm]
Figure 5: Relationship between mean tree SID and ring shake index.
70
80
90
100
102
Section B
6a) Bedigliora 14
SID [ µ m]
200
150
8
150
8
2
1
SID [ µ m]
200
2
150
8
0
3
6
7
4
1
8
7
1
8
3
4
1
8
3
4
5
4
5
1
2
7
8
3
6
4
5
3
0
6
4
2
6
7
5
7
5
50
3
6
5
2
100
0
4
6
2
6
7
1
150
8
2
50
3
0
5
200
100
50
50
6d) Brione 9
SID [µ m]
200
100
100
7
6c) Bedigliora 5
1
1
SID [µ m]
6b) Brione 11
2
7
3
6
4
5
Figure 6: Example of different patterns of ring shake incidence and SID distribution in the outer part of
the trunk. 6a) low ring shake incidence by high SID value; 6b) high ring shake incidence by low SID
value; 6c) high ring shake incidence in correspondence to a disc sector displaying high SID value; 6d)
high ring shake incidence in correspondence to a disc sector displaying lower SID value.
General consideration
Longitudinal surface SID measurements showed that there is a large intra-tree-variation
in one single tree as well as a large inter-tree-variation within trees belonging to the
same stand. This is quite normal for tree- and wood-quality parameters, which typically
depend on several internal and external variables. Nevertheless, an overall comparison
of the SID between stands shows a tendency towards a lower SID value in
correspondence to the stand slope direction. This observation is likely associated with
the effect that tension wood has on tensile stresses (Kubler, 1987). The silvicultural
treatment appears not to have affected mean SID values the stand. A slight increase in
SID values was however observed in correspondence of wider annual growth rings.
Nevertheless, this study does not reveal any clear relationship between ring shake
occurrence and the longitudinal surface SID. In both stressed chestnut trees or in
stressed stem sectors no increased incidence of ring shake was found. However there are
several methodological limits that could have masked the existence of this relationship,
independent of the radial strength capacity of a single chestnut individual. On one hand,
the single hole drilling method performed on standing trees permits us to indirectly
characterize growth stresses that affect solely the outer part of the stem. Because the
SID is measured as a result of a 20 mm deep hole, it is not clear how stresses which are
“hidden” inside the stem, influence the development of ring shake, which most often
appears inside the stem (Macchioni and Pividori, 1996; Fonti et al., 1998). On the other
hand, in order to characterize the effective stress value, the deformation observed has to
be multiplied by the modulus of elasticity, which could differ among the measurement
points and therefore partially cover the relationship between SID and stress.
On the basis of current observations it is therefore not possible to assess or reject
whether ring shake incidence is related to high growth stress levels.
Section B
103
Acknowledgements
A very special thank you goes to Franco Fibbioli for his help with field work and to
Robert Widmann for his precious advice. I also thank Jürgen Sell and Marco Conedera
for comments and advice concerning this manuscript.
References
Archer, R. (1986). Growth stresses and strain in trees, Springer Verlag Berlin
Elzière, S. (1995). “Influence de l'éclarcie sur des brins de taillis de châtaignier (Castanea sativa Mill.)
Analyse de tige, contraintes de croissance et fissilité du bois,” Mémoire, Ecole national d'ingénieur
des travaux agricoles de Bordeaux, Bordeaux, 57.
EU-FAIR-Project (2001). "Stresses in beech: occurrence and relevance of growth stresses in beech
(Fagus silvatica L.) in central Europe". Final report, FAIR-Project CT98-3606, 5 Framework
Program, 323.
Fonti, P., Giudici, F., Kucera, L. J., Ott, E., and Pöhler, E. (1998) “Studio sulla cipollatura in un ceduo
castanile.” Atti del convegno nazionale sul castagno, Cison di Valmarino (Treviso), 293-302.
Fonti, P., and Macchioni, N. (submitted). “Ring shake in chestnut: anatomical description, extent and
frequency of failures.” Annals of forest science.
Fonti, P., Macchioni, N., and Thibaut, B. (2002). “Ring shake in chestnut (Castanea sativa Mill.): state of
the art.” Annals of Forest Science, 59(2), 129-140.
Fournier, M., Chanson, B., Thibaut, B., and Guitard, D. (1994). “Mesures de déformation résiduelles de
croissance à la surface des arbres en relation avec leur morphologie. Observations sur différentes
espèces.” Annals of Forest Science, 51, 249-266.
Kubler, H. (1987). “Growth stresses in trees and related wood properties.” Forest Products Abstracts,
10(3), 61-119.
Macchioni, N., and Pividori, M. (1996). “Ring shake and structural characteristics of a chestnut
(Castanea sativa Mill.) coppice stand in northern Piedmont (northwest Italy).” Annals of Forest
Sciences, 53, 31-50.
Thibaut, B. (1994). “New silvicultural methods and innovative technologies for the valorisation of
Chestnut wood as a prime resource for industry.” Contract MA2B-CT91-0027, Forest Program
Chestnut - Task C - Operative Unit 9, 21.
Thibaut, B., Fournier, M., and Jullien, D. (1995). “Contraintes de croissance, recouvrance différée à
l'étuvage et fissuration des grumes: cas du châtaignier.” Forêt Méditerranéenne, 16(1), 85-91.
CURRICULUM VITAE
Personal details:
Family name
Name
FONTI
Patrick
Date of birth
Nationality
Place of origin
14.09.1971
Switzerland
Miglieglia (TI)
Education:
June 1997-1999
1992-1997
1986-1991
Postgraduate course in applied statistic at the Swiss Federal
Institute of Technology in Zürich
Diploma in Forest science, Faculty of Forest sciences,
Swiss Federal Institute of Technology in Zürich
Scientific Matura, High school in Lugano (CH)
Professional experiences:
2000
June 1997-1999
Research assistant at the Swiss Federal Institute of
Technology in Zurich (CH)
Project collaborator at the WSL Swiss Federal Institute for
Forest, Snow and Landscape Research in Bellinzona (CH)
Current activity:
2000-2002
Doctoral student at the Swiss Federal Institute of
Technology in Zurich (CH) and at the WSL Swiss Federal
Institute for Forest, Snow and Landscape Research in
Bellinzona (CH).
SUMMARY
This work is concerned with the development of ring shake, a wood defect
that very often occurs in European chestnut (Castanea sativa Mill.). This
study must be placed in a general research context aimed at assembling a
better picture of the complex phenomenon of ring shake in order to
evaluate new preventive measures that will minimise the risk of
occurrence. The main objectives are (1) the examination of the
relationships between radial wood cohesion, growth stresses and ring shake
occurrence, and (2) an analysis of the effect that improved growth due to
new silvicultural concepts aimed at quality coppice wood production has
on ring shake occurrence.
Investigative studies were carried out on wood material collected in
chestnut coppice forests from the southern part of the Swiss Alps. Analyses
were performed on the basis of a comparison of wood with and without
ring shake.
Anatomical investigations of the ring shake privileged plane of fracture
(earlywood area) showed that ring shaken wood displays a higher amount
of rays and smaller earlywood vessel lumina. Mechanical rupture testing
revealed that earlywood radial split resistance is positively correlated with
ring width. Examination of longitudinal strain-induced displacement
performed on the stem surface of coppice trees did not reveal the existence
of a direct relationship between high stress level and ring shake occurrence.
Lastly, through a review of literature combined with a description and
quantification of the different typologies of chestnut ring shake so far
observed, important elements for the interpretation of results were
collected.
It has been highlighted that the principal cause of ring shake in chestnut is
associated with the unbalanced interrelation between the weak structure of
the tangential wood plane and the release of growth stresses. Stress
overloaded trees react by producing a strengthened wood. When trees are
cut however, mechanically originated ring shakes develop as a
consequence of the relieving of growth stresses. Nevertheless mechanical
testing proved that not all trees are prone to ring shake in the same way.
This result is particularly encouraging because, compatible with the new
silvicultural methods, it is then possible for the forester to operate in a
manner which reduces the risk of ring shake.

Introduction

Transcription

Documents pareils

Motion Technology Slip Ring Catalog

sustainable treatments (BioCoPol) - WoodWisdom-Net

DISTRIBUTION OF YOUNG TREES The Ministère des Forêts, de la

Vilac Holiday Preview 2016

Ouverture Geant Casino Aix En Provence

RESOLUTE FOREST PRODUCTS Investor Presentation

Date : July 2015

ETTF News Winter 2014 - The European Timber Trade Federation

fifth international conference

Appendix 2

Revision of Green Public Procurement Criteria for furniture.