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 ring shake occurrence 60 wood discs from 3 mature coppice stands (Novaggio, Gerra, Bedano). 10 shake-free and 10 with extreme ring shake occurrence from each selected stand. 60 wood discs from 3 mature coppice 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 ring shake occurrence 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. Annals of forest science (2002), Vol. 59 (2), 129-140 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]. Ring shake in chestnut: state of the art 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 Observation at the base of the logs 42 - [18] 156 shoots taken from 6 regions Observation at the base of the 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 Observation on 5 cm thick disks taken at different heights, starting from the base of the logs Observation on 5 cm thick disks taken at different heights, starting from the base of the logs Observation on 5 cm thick disks taken from the bases of 300 shoots Observation on 5 cm thick disks taken at different heights, starting from the base of the logs Observation on 5 cm thick disks 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]. Annals of forest science (2002), Vol. 59 (2), 129-140 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]. Ring shake in chestnut: state of the art 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 Annals of forest science (2002), Vol. 59 (2), 129-140 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 Ring shake in chestnut: state of the art 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). Annals of forest science (2002), Vol. 59 (2), 129-140 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 Ring shake in chestnut: state of the art 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. Annals of forest science (2002), Vol. 59 (2), 129-140 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. Ring shake in chestnut: state of the art 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. Annals of forest science (2002), Vol. 59 (2), 129-140 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. Ring shake in chestnut: state of the art 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. Annals of forest science (2002), Vol. 59 (2), 129-140 Section B 45 2 Ring shake in chestnut: anatomical description, extent and frequency of failures Patrick Fontia and Nicola Macchionib 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 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. Annals of forest science, submitted 2002 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, Ring shake typologies 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. Annals of forest science, submitted 2002 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 Year of the annual ring Figure 6: Ring shake distribution. Ring shake typologies 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. Annals of forest science, submitted 2002 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. Ring shake typologies 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]). Annals of forest science, submitted 2002 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. Ring shake typologies 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. Annals of forest science, submitted 2002 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. Relationship ring shake – earlywood vessels 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. Relationship ring shake – earlywood vessels 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. Relationship ring shake – earlywood vessels 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. Relationship ring shake – earlywood vessels 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). Relationship ring shake – earlywood vessels Section B 77 Results and discussion 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 Trees, structure and function (2002) , Vol. 16 (8), 519-522 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. Relationship ring shake – earlywood vessels 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 Ray volume fraction [%] 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% Ray volume fraction [%] 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. Trees, structure and function (2002) , Vol. 16 (8), 519-522 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 Relationship ring shake – earlywood vessels 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 Trees, structure and function (2002) , Vol. 16 (8), 519-522 Section B 83 5 Radial split resistance of chestnut earlywood and its relation to the ring width Patrick Fontia and Jürgen Sellb a WSL Swiss Federal Research Institute, Sottostazione Sud delle Alpi, Via Belsoggiorno 22, Casella postale 57, CH-6504 Bellinzona, Switzerland 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. Materials and methods 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. Wood and fiber science, in press 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; Radial split resistance and ring width in chestnut 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. Wood and fiber science, in press 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 Year of the annual ring 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. Radial split resistance and ring width in chestnut 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). Wood and fiber science, in press 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 Radial split resistance and ring width in chestnut 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. Wood and fiber science, in press Section B 95 6 Growth strain and ring shake in chestnut coppice trees Patrick Fontia a WSL Swiss Federal Research Institute, Sottostazione Sud delle Alpi, Via Belsoggiorno 22, Casella postale 57, CH-6504 Bellinzona, Switzerland 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 Materials and methods 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. Results and discussion 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. Forêt méditerranéenne, submitted 2002 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. Growth strain and ring shake incidence 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. Forêt méditerranéenne, submitted 2002 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 Ring shake incidence 1 2 7 8 3 6 4 5 3 0 6 4 Ring shake incidence 2 6 7 5 7 5 50 3 6 5 Ring shake incidence 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 Ring shake incidence 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. Growth strain and ring shake incidence 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. Forêt méditerranéenne, submitted 2002 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.