METSÄNTUTKIMUSLAITOKSEN TIEDONANTOJA 880,2002 FINNISH FOREST RESEARCH INSTITUTE, RESEARCH PAPERS 880, 2002 Decay resistance of heartwood timber as a quality characteristic in Scots pine breeding Martti Venäläinen PUNKAHARJUN TUTKIMUSASEMA - PUNKAHARJU RESEARCH STATION METSÄNTUTKIMUSLAITOKSEN TIEDONANTOJA 880, 2002 FINNISH FOREST RESEARCH INSTITUTE, RESEARCH PAPERS 880, 2002 Decay resistance of heartwood timber as a quality characteristic in Scots pine breeding Martti Venäläinen Akateeminen väitöskirja metsänjalostuksen alalta Esitetään Helsingin yliopiston maatalous-metsätieteellisen tiedekunnan luvalla julkisesti tarkastettavaksi Suomen metsämuseo ja metsätietokeskus Luston auditoriossa, 23. päivänä tammikuuta 2003 kello 12. Academic dissertation in forest tree breeding To be presented with the permission of the Faculty of Agriculture and Forestry of the Uni versity of Helsinki, for public criticism in the auditorium of Lusto, at Punkaharju, on the 23rd of January 2003, at 12 noon. Piorum ac proborum censurae in vicem publici sped minis. PUNKAHARJUN TUTKIMUSASEMA - PUNKAHARJU RESEARCH STATION 2 Decay resistance of Scots pine heartwood Venäläinen, Martti 2002. Decay resistance of heartwood timber as a quality characteristic in Scots pine breeding. Seloste: Sydänpuutavaran lahonkestävyys laatuominaisuutena männynjalostuksessa. Metsäntutkimuslaitoksen tiedonantoja 880. Finnish Forest Research Institute, Research Papers 880. Academic disser tation. 53 + 69 p. ISBN 951-40-1866-4. ISSN 0358-4283. Vastaväittäjä Doc. Tapani Pöykkö Opponent Häme Polytechnic, Evo Esitarkastajat Dr. Björn Hannrup Reviewers Skogforsk, Uppsala Science Park, Sweden Dr. John H. Russell Cowichan Lake Research Station, Canada Työn ohjaajat Doc. Anni Harju Supervisors Finnish Forest Research Institute, Punkaharju Dr. Hannu Viitanen VTT Transport and Building, Espoo Kirjoittajan yhteystiedot Finnish Forest Research Institute Author's address Punkaharju Research Station Finlandiantie 18 FIN -5 8450 Punkaharju tel. + 358 15 7302 238, fax: +358 15 644 333 e-mail: martti.venalainen@metla.fi Julkaisija Metsäntutkimuslaitos, Punkaharjun tutkimusasema Publisher Finnish Forest Research Institute, Punkaharju Research Station Accepted to be published by Eeva Korpilahti, editor-in-chief, 2.12.2002 Martti Venäläinen 3 Abstract The aim of this study was to evaluate the biological possibility of improving the decay resistance of Scots pine (Pinus sylvestris L.) heartwood by means of forest tree breeding. The two fundamental prerequisites for successfully carrying out this task were: a method suitable for testing the decay resistance of standing trees, and tree material suitable for analysis of the genetic pa rameters of heartwood characteristics. A satisfactory method for investigating the natural decay resistance of the wood of standing trees proved to be the in vitro accelerated malt agar plate decay test, which uses increment core sec tions as samples. The decay tests, carried out with two relatively young, 32- 34 year-old, Scots pine progeny test materials, showed that there was pheno typic variation in the decay resistance of juvenile heartwood among the indi vidual trees. In the two progeny tests, the coefficient of additive genetic variation was estimated to be CVA = 10.6 % and CVA = 28.5 %, and the pro portion of the additive genetic variation out of the total phenotypic variation, i.e. the individual heritability, was estimated to be h 2 = 0.02 and h 2 = 0.37. It was concluded that the additive genetic variation was adequate for a consid erable response to phenotypic selection in the decay resistance of juvenile heartwood. However, the decay tests with mature, 170-year-old, Scots pines showed the most durable part to be the outer heartwood. In practical tree breeding direct selection for the durability of outer heartwood would prolong the generation interval unrealistically. In order to elucidate the possibilities of indirect selection, the relationships between decay resistance and other chemical or physical wood characteristics were studied in the progeny test material as well as in a mature natural stand. The most promising characteristic for indirect selection appeared to be the concentration of two secondary phenolic compounds, pinosylvin (PS) and pinosylvin monomethyl ether (PSM), both of which are stilbenes. The resin acids, which are important for the active defence of living trees, seemed to play such a weak role in passive defence that their concentration does not provide any useful information for indirect durability selection. The stilbenes, which are normally synthesised in heartwood formation when the last living sapwood cells die, but which evidently can be induced by artificial stress fac tors to form already in young seedlings, could even provide a tool for an early indirect selection. However, before any indirect selection will be per formed, it is advisable to carry out a thorough investigation of the genetic correlations between the characteristics used in indirect selection and the tar get characteristic as well as the growth and quality characteristics already in cluded in the Scots pine breeding programme. Keywords: Pinus sylvestris, juvenile heartwood, genetic parameters, early indirect selection, Coniophora puteana, accelerated decay test, resin acids, stilbenes, pinosylvin, water sorption Decay resistance of Scots pine heartwood 4 Seloste Sydänpuutavaran lahonkestävyys laatuominaisuutena männynj alostuksessa Puuaineen luontainen lahonkestävyys johtuu ensisijaisesti sydänpuu hun muodostuvista lajikohtaisista uuteaineista, jotka eri tavoin haittaavat lahottajasienten hajotustoimintaa. Lahonkestävyys vaihtelee sekä puulaji en että rungon eri osien välillä. Ihmiset ovat vuosituhansien myötä oppi neet käyttämään tätä vaihtelua hyväkseen ja valitsemaan puutavaran käyt töolosuhteiden mukaisesti. Nykyinen kiinnostus luontaiseen lahonkestä vyyteen on virinnyt sitä mukaa kuin kyllästettyä puutavaraa koskevat vi ranomaismääräykset ovat tiukentuneet ja kuluttajien asenteet kyllästys kemikaaleja kohtaan ovat muuttuneet varovaisemmiksi. Meillä Suomessa männyn sydänpuu voisi olla kuluttajien arvostama kotimainen ja ekologi sesti hyväksyttävä puutavaralaji sellaisiin kosteudelle alttiisiin rakennus kohteisiin, joissa ei edellytetä ehdotonta lahonkestävyyttä, mutta joissa toivotaan kohtuullisen pitkää käyttöikää. Tällaisia ovat esimerkiksi ne pi harakenteet ja laiturien osat, jotka eivät ole jatkuvassa maa- tai vesikoske tuksessa. Tämän väitöskirjatyön tarkoituksena on ollut selvittää ja pohtia biologiselta kannalta, voidaanko männyn sydänpuun lahonkestävyyttä pi tää metsänjalostuksessa varteenotettavana laatuominaisuutena. Väitöskirjatyön empiiriset havainnot on julkaistu kuudessa tutkimus artikkelissa. Artikkelissa I kuvataan kasvukairalla pystypuusta otettuun sydänpuunäytteeseen perustuva pikalahotusmenetelmä. Tällainen tarvit tiin, koska puutavaran lahonkestävyyden arvioimiseen kehitetyt stan dardimenetelmät, jotka vaativat puun kaatamista, eivät sovellu metsän jalostustutkimuksiin. Tutkimuksissa Ilja 111 tarkasteltiin kahdesta (32- ja 34-vuotiaasta) männyn jälkeläiskokeesta peräisin olevan, sydänpuuksi muuttuneen nuorpuun ominaisuuksien vaihtelua. Tarkastelu kohdistui erityisesti lahonkestävyydessä havaitun additiivisen geneettisen vaihtelun määrään suhteessa lahonkestävyyden keskiarvoon ja fenotyyppisen vaihtelun kokonaismäärään. Tutkimuksissa IV ja V tarkasteltiin, poikkea vatko hitaasti ja nopeasti lahoava sydänpuu toisistaan hartsihappojen tai fenoliyhdisteiden pitoisuuksien suhteen. Nämä molemmat uuteaine ryhmät ovat tyypillisiä männyn sydänpuulle. Tutkimuksessa VI tarkas teltiin lahoamisnopeuden ja puuaineen kemiallisten ja fysikaalisten omi naisuuksien välisiä riippuvuuksia 170-vuotiaissa luonnonmetsän puissa. Kairanlastunäytteestä tehty pikalahotuskoe antoi tarkkuudessaan tyy dyttävän tuloksen pystypuun lahonkestävyydestä, joten jalostusaineistoon kuuluvien, elävinä säilytettävien puiden testaaminen todettiin mahdolli seksi (I). Nuorten mäntyjen sydänpuun lahonkestävyyden vaihtelu havaittiin varsin laajaksi, mikä puolestaan tekee fenotyyppisen valinnan mahdolliseksi. Lahonkestävyyden additiivisen geneettisen vaihtelun Martti Venäläinen 5 määrä suhteessa keskiarvoon havaittiin kummassakin jälkeläiskokeessa siinä määrin korkeaksi, CVa = 10.6 % (II) ja CVA = 28.5 % (III), että vanhempien valinnalla voidaan olettaa saavutettavan merkittävää vastetta jälkeläisten keskiarvossa. Additiivisen geneettisen vaihtelun ja feno tyyppisen kokonaisvaihtelun suhteesta eli heritabiliteetista saatiin kaksi erisuuruista estimaattia. Korpilahden jälkeläiskokeessa havaittu korkea heritabiliteetti, h 2 = 0.37 (HI), viittasi siihen, että yksilöiden fenotyyp pinen arvo korreloi merkittävässä määrin jalostusarvon kanssa. Kerimäen jälkeläiskokeessa heritabiliteetti oli alhainen (h 2 = 0.02, II). Tuloksista pääteltiin, että lahonkestävyyden additiivinen geneettinen vaihtelu on jalostuksellista valintaa ajatellen riittävä, ja että Korpilahden jälkeläiskoe edustaa yksilövalintaan sopivaa kohdetta. Lahonkestävyyden suora varhaisvalinta manto- eli pintapuun perus teella todettiin hyödyttömäksi (II). Suora valinta on mahdollista vasta sitten, kun mänty alkaa ilmentää lahonkestävyysominaisuuttaan. Tämä tapahtuu noin 30-vuotiaana, kun sydänpuun muodostuminen rungon keskellä olevassa nuorpuussa on päässyt vauhtiin (11, III). Vanhoista luonnonmännyistä todettiin, että sydänpuun ulompi osa oli merkittävästi kestävämpää kuin sisäosan nuorsydänpuu (VI). Mikäli jalostuksellisessa valinnassa jäätäisiin odottamaan kypsän mantopuun muuttumista sydän puuksi, mikä merkinnee vähintään 50 vuotta, sukupolvien väli venyisi jalostuksen etenemisen kannalta epäedullisen pitkäksi. Lahonkestävyyden epäsuoran mittaamisen ja valinnan mahdollisuuk sien kannalta lupaavimmaksi ominaisuudeksi osoittautui fenoliyhdistei siin kuuluvien stilbeenien (pinosylviini ja pinosylviinimonometyylieette ri) pitoisuus (V, VI). Elävän puun aktiivisen puolustautumisen kannalta tärkeiden hartsihappojen määrän vaikutus passiiviseen lahonkestävyyteen todettiin vähäiseksi (IV, VI). Stilbeenien tiedetään normaalisti syntyvän sydänpuuksi muuttumassa olevan mantopuun ydinsäteiden tylppysolujen, ts. viimeisten elävien solujen, kuollessa hitaasti. Kirjallisuuden perusteel la todettiin, että stressitekijöiden (esim. UV-säteilyn tai mekaanisen vau rioittamisen) avulla stilbeenien synteesi saadaan tapahtumaan jo pienissä taimissa, mikä saattaa mahdollistaa lahonkestävyyden epäsuoran varhais valinnan. Epäsuoran valinnan mahdollisuuksien selvittäminen vaatii pe rusteellisia jatkotutkimuksia erityisesti valinnan kohteena olevien ominai suuksien, lahonkestävyyden ja männyn jalostusohjelmassa jo mukana olevien kasvu-ja laatuominaisuuksien geneettisistä korrelaatioista. Väitöskirjan loppupäätelmissä ehdotetaan männyn jalostuksen paino pisteen siirtämistä kasvunopeudesta rungon ulkoisten ja puuaineen sisäis ten laatuominaisuuksien suuntaan. Avainsanat: mänty, kellarisieni, nuorpuu, sydänpuu, lahotustesti, epä suora valinta, periytyvyys, pihka, hartsihappo, fenoliyhdiste, pinosylviini 6 Decay resistance of Scots pine heartwood Alkusanat Tämä väitöskirja on metsänjalostuksen näkökulmasta tehty yhteenveto niistä tutkimuksista, jotka tehtiin vuosina 1998-2002 Metsäntutkimuslaitoksen tutkimushankkeen 3220, "Suomalaisten puulajien luontainen käyttökelpoisuus rakentamisessa", osahankkeen 01, "Männyn ja lehtikuusen perimän vaikutus puun lahonkestävyyteen" sekä Wood Wisdom metsäalan tutkimusohjelmaan sisältyneen Suomen Akatemian rahoittaman tutkimushankkeen 43140, "Männyn ja siperianlehtikuusen puuaineen lahonkestävyyden geneettinen vaihtelu" saumattomana yhteistyönä. Kuten alkuperäisten julkaisujen luettelosta käy hyvin ilmi, tutkimukset ovat olleet samalla usean eri tieteenalan, usean tutkimuslaitoksen ja ennen kaikkea usean tutkijan yhteistyötä. Olen kiitollinen teille Anni Harju, Teijo Nikkanen, Leena Paajanen, Pirkko Veiling, Hannu Viitanen, Egbert Beuker, Pirjo Kainulainen, Markku Tiitta, Pekka Saranpää ja Hanna Nikulainen kanssakirjoittajina sopuisasta ja tunnollisesta uurastuksesta niin havaintoaineistojen hankkimisessa ja käsittelyssä kuin raporttien valmistelussakin. Kaikkiaan olen kiitollinen sille monien vaiheiden ketjulle, joka on vähitellen avannut minulle kiehtovan ikkunan puun ja sen uuteaineiden ominaisuuksiin ja lahottajasienten maailmaan. Opinnäytetyötäni ovat ohjanneet Anni Harju ja Hannu Viitanen. Hannulta perin sen hankeratsun ohjakset, joka vei urakelkkani jalostusohjelmista ja puun ulkoisen laadun tutkimuksesta keskelle lahonkestävyyden tutkimisen hetteikköä. Anni tuli "lahohankkeeseen" pätevöitymään metsägenetiikan dosentiksi ja mitä geneettisten parametrien estimoimisen opettamiseen tulee, sai harjoitusmateriaalikseen kyllin haastavan oppilaan. Kiitän sydämellisesti teitä kumpaakin omasta kontribuutiostanne työn aikana ja käsikirjoituksen korjailussa. Oheiset tutkimukset eivät olisi valmistuneet koskaan ilman poikkeuk sellisen mittavaa avustajakaartia. Esko Oksa, Pentti Manninen, Reijo Rauniomaa, Hannu Partti, Tapani Relander, Outi Kylliäinen, Markku Ti ainen, Liisa Seppänen, Heikki Paajanen, Heikki Kinnunen, Teemu Moi lanen, Eija Matikainen, Tore Ericsson, Matti Haapanen, Seppo Ruotsalai nen, Tarja Salminen, Jussi Tiainen, Arto Haatainen, Terhi Vuorinen, Ta pio Laakso, Veikko Kitunen, Ari Haapasaari, Pentti Konttinen, Hannu Heinonen, Sari Lignell, Katriina Lipponen, Anna-Maija Hallaksela, Kert tu Rainio, Auvo Silvennoinen, Hannele Makkonen ja Seija Vatanen, osuuttanne on kuvattu suppeasti alkuperäisten julkaisujen yhteydessä. Kaikille teille ja niille nimeltä mainitsematta jääneille Punkaharjun, Van taan ja Joensuun yksikköjen henkilöille, jotka ovat tutkimuksiamme Martti Venäläinen 7 eteenpäin vieneet, lausun vilpittömät kiitokseni. Ulotan kiitokseni koske maan myös huolto- ja toimistohenkilökuntaa, sekä yksiköiden päällikkö jä, tutkimuksen asiantuntijaryhmää että Metlan koko esikuntaa. Two qualified scientists, nominated as official reviewers by Helsinki university, Björn Hannrup from Sweden and John H. Russell from Can ada worked thoroughly but rapidly in evaluating the thesis. I extend cor dial thanks for you for your constructive contribution! Käsikirjoituksen ovat lukeneet Matti Haapanen, Piijo Kainulainen ja Pirkko Veiling. Kiitän teitä arvokkaista kommenteistanne ja siitä, toimi titte ne minulle eikä vastaväittäjälle. Oman erityiskiitoksensa ansaitsee kielieditori John Derome, joka on sinnikkäästi koijannut samoja heik kouksia kaikista osajulkaisuista ja lopulta yhteenvedosta. Eija Matikaisel le kiitos kirjan taittamisesta. Loppulauseeni osoitan maailman parhaille tyttärille Sirjelle, Pinjalle, Pihlalle ja Milkalle. Moni asia, jota olette kärsivällisesti pyytäneet, on jäänyt viime aikoina tekemättä. Toivon vähitellen voivani korvata takai sin sen poissaolon, jonka tämän väitöskirjan kirjoittaminen on aiheutta nut. Punkaharjulla, joulukuun 12. päivänä 2002 Martti Venäläinen 8 Decay resistance of Scots pine heartwood Original publications The thesis is based on the following publications, which are referred to in the text by Roman numerals. All the publications are reprinted with the permission of the copyright holders. I Venäläinen, M., Harju, A., Nikkanen, T., Paajanen, L., Veiling, P., & Viitanen, H. 2001. Genetic variation in the decay resistance of Sibe rian larch (Larix sibirica Ledeb.) wood. Holzforschung 55(1): 1-6. II Harju, A.M., Venäläinen, M., Beuker, E., Veiling, P., & Viitanen, H. 2001. Genetic variation in the decay resistance of Scots pine wood against brown rot fungus. Can. J. For. Res. 31(7): 1244-1249. 111 Harju, A .M. & Venäläinen, M. 2002. Genetic parameters regarding the resistance of Pinus sylvestris heartwood to decay caused by Co niophora puteana. Scand. J. For. Res. 17:199-205. IV Harju, A.M., Kainulainen, P., Venäläinen, M., Tiitta, M., & Viitanen, H. 2002. Differences in resin acid concentration between brown-rot resistant and susceptible Scots pine heartwood. Holzforschung 56(5):479-486. V Venäläinen, M., Harju, A.M., Saranpää, P., Kainulainen, P., Tiitta, M., and Veiling, P. 2003. The concentration of phenolics in brown-rot de cay resistant and susceptible Scots pine heartwood. Wood Sci. Tech nol. In press. VI Venäläinen, M., Haiju, A.M., Kainulainen, P., Viitanen, H. and Niku lainen, H. 2003. Variation in the decay resistance and its relationship with other wood characteristics in old Scots pines. Ann. For. Sci. In press. The contribution of Martti Venäläinen in the preparation of the original articles: 11, VI main responsibility for planning the experiment, collect ing the material and handling the samples I, V, VI main responsibility for data analysis and writing the manuscripts 111, IV, V joint responsibility for planning the experiment, collecting the material and handling the samples 11, 111, IV joint responsibility for data analysis and writing the manuscripts Martti Venäläinen 9 Contents Abstract 3 Seloste 4 Alkusanat 6 Original publications 8 Contents 9 1. Introduction 10 1.1 Scots pine breeding guided by the ideotype 10 1.2 Need to estimate genetic parameters 11 1.3 The dilemmas of simultaneous selection for several characteristics and early selection 13 1.4 The current criteria of Scots pine wood quality 13 1.5 Natural durability 14 1.5.1 The 'rehabilitation' of natural durability 14 1.5.2 Factors affecting natural durability 15 1.5.3 In vitro methods for studying natural durability 17 1.6 Aim of the study 18 2. Material and methods 19 2.1 Wood material 19 2.2 Decay tests 20 2.3 Quantification of extractives 21 2.4 Water sorption capacity 22 2.5 Statistical analysis 22 3. Results and discussion 25 3.1 Measuring the decay resistance of timber from standing trees 25 3.2 Phenotypic variation in the decay resistance of Scots pine 26 3.3 Additive genetic variation of decay resistance in sapwood and juvenile heartwood 27 3.4 Quantity of heartwood 28 3.5 Prospects of direct selection 29 3.6 Prospects of indirect selection: phenotypic relationships between decay resistance and other wood characteristics 31 3.6.1 Phenotypic versus genetic correlation 31 3.6.2 Resistance gatework 31 3.6.3 Water repellency 33 3.6.4 Resin acids 34 3.6.5 Stilbenes 34 3.6.6 Other factors 36 3.7 Genetic correlations 37 3.8 The role of environment 38 3.8.1 Selection environment 38 3.8.2 The role of environment in the growing of improved Scots pines 39 4. Conclusions 41 5. References 44 Appendix 52 10 Decay resistance of Scots pine heartwood 1. Introduction 1.1 Scots pine breeding guided by the ideo type Tree breeding is an essential part of intensive forestry, which is based on the regeneration of stands through cultivation and on the active manage ment of the stands during their growth. In addition to the natural condi tions, such as soil, climate, and the risk of damage, the demands of the wood markets direct the forest owners to decide how the stands should be regenerated and managed in order to maximise their economic return. Al though the quantity of wood produced will always be an important target of forestry, the quality of wood seems to become an ever more important factor in the wood markets (Paavilainen 2002). The tree breeders have to be far ahead of the forest owners when searching for signs of what kind of wood material will be considered valuable in the future. The improved products of the tree breeders, i.e. cultivars, have to survive and grow vig orously for decades and produce a large amount of high quality timber. In addition, it would also be advantageous for the cultivars to have a positive interaction with the management practices, i.e. that the stands would 'pay back with a high rate of interest' the investments made in intensive man agement aimed at enhancing either volume production or a specific qual ity property of the wood. In order to integrate all the requirements set on improved trees, the concept of an ideotype, which Donald (1968) has introduced to crop plant breeding, has been adopted and developed to tree breeding (Dickmann 1985, Dickmann et al. 1994). Today the ideotype of the cultivated tree can be seen as an ideal model of a tree that meets several, if not all, of the requirements set on the cultivar as regards the yield and behaviour in a specific environment. Thus the adoption of the ideal model as a breeding goal provides an important practical tool for breeding strategy planning. For Scots pine breeding in Finland the ideotype has been delineated by Kärki and Tigerstedt (1985), Veiling (1988) and Pöykkö (1993). High volume growth on an areal basis, combined with good external quality for the sawing industry, have been the key components of the Scots pine ideotype. Good external quality is associated with thin branches in rela tion to the diameter of the stem, and thus the ideotype also includes the target of high harvest index, i.e. a high yield of harvested timber com pared to the total biomass of the stem (Donald 1962, Tigerstedt and Vei ling 1986). In actual fact, the ideotype was intuitively applied already in the phenotypic selection of plus trees (Oskarsson 1995), although the concept itself was not familiar to the tree breeders at that time. Martti Venäläinen 11 The requirements set on a cultivar should not represent a biological contradiction, e.g. simultaneous extractive-rich and extractive-free heart wood. Otherwise more than one ideotype is needed. A comprehensive physiological and genetic investigation is always needed within each ideotype, in order to decide whether the breeding efforts can be directed simultaneously at all the components of the ideotype. In the implementation of the Finnish breeding programme for Scots pine, recurrent selection of the plus trees has so far been mainly based on the height growth of the progenies (Venäläinen and Ruotsalainen 2002). In terms of Hannrup (1999), tree height has been used as "an operational selection criterion" for the target trait, "the total dry matter production", as has also been the case in Sweden. This is partly due to the fact that the quality characteristics are manifested at a later age in the progeny tests, and the costs of quality measurements limit the amount of data that can be collected. In order to simplify the quality measurements and analysis of the data, visual assessment of young trees has been proposed (Venäläinen et ai. 1996). On the other hand, Haapanen et ai. (1997) recently stated that the total height has been a lucky choice as a selection criterion because it appears to have a favourable genetic correlation with the external quality characteristics. 1.2 Need to estimate genetic parameters The genetic parameters of a quantitative characteristic in a certain popula tion include the additive genetic variance, the heritability, and the genetic correlation between pairs of characteristics. It would be beneficial to have estimates for these parameters in the populations with which the tree breeders operate: natural stands in the early phases of selection, progeny tests at a later stage and perhaps the virtual breeding populations in the advanced phases of breeding. In practice, in the initial phase of tree breeding programmes, plus tree selection is started with no information about the genetic parameters if the phenotypic variation is considered to be sufficient. However, the existence of additive genetic variation is the prerequisite for the selection response in the next generation. On the other hand, the statistical requirement for estimation the genetic variance com ponents is the relatedness of the individuals used in the analysis (Falconer 1981). As a result, estimation is not possible before the progeny tests have been established. This is the rationale on which the traditional tree breeding procedure is based. The additive genetic variance is, as such, a relatively abstract parame ter. Analogous to the coefficient of phenotypic variation (CV %), defined as the ratio of the phenotypic standard deviation and the phenotypic 12 Decay resistance of Scots pine heartwood mean, is the coefficient of additive genetic variation, CV A %, which is defined as the ratio of the additive genetic standard deviation, oA, and the phenotypic mean (Houle 1992, Cornelius 1994). Thus the estimate of CV A helps us to evaluate the absolute scale of the gain, i.e. the magnitude of the selection response compared to the current average of the charac teristic in the population. The individual heritability, h 2 , i.e. the ratio between the additive ge netic variance and the phenotypic variance, expresses how precisely the genotypic value of the individual tree can be predicted from its phenotype in the analysed population (Falconer 1981). If the estimate of the herita bility is low, then the phenotype mainly reflects environmental effects. Thus in the tree breeding context, the heritability of a characteristic, esti mated in a specific progeny test, tells us how good the progeny test is as a selection environment for the characteristic in question (White and Hodge 1989). The average level of several heritability estimates is interpreted to depict the "strength of genetic control" and thus "the possible gains through selection and breeding" for the characteristic in question, e.g wood density (Zobell and Jett 1995). Estimates of the genetic correlation between characteristics are neces sary before the recurrent selection can be started. If indirect selection is to be applied, the genetic correlations enable the breeder to predict the response of indirect selection in the target characteristic. In the case of index selection, the genetic correlations are essential elements when con structing a selection index (Cotterill and Dean 1990). A genetic correla tion can also reveal an 'indirect response' of selection in a characteristic that is not supposed to change. The heritability and the CV A together provide prospects of improving the average value of a characteristic through selection. Both estimates are also needed to determine which is the most efficient breeding strategy for each characteristic (Hannrup 1999). In the planning of a breeding pro gramme it might be advisable to set the different components of the ideo type into a 2 x 2 table (Table 1) as soon as the necessary parameters are available. Table 1. Prospects of improving a characteristic by breeding, predicted on the basis of the scale of the heritability and the coefficient of the additive genetic variation. Modified from Hannrup (1999). High CVA Low CV A H ■ easy breeding progress only by intensive selection Low h 2 progress only through intensive testing difficult breeding Martti Venäläinen 13 1.3 The dilemmas of simultaneous selection for several characteristics and early se lection Even a fairly simple ideotype leads to a breeding task in which the gain in the next generation should be achieved in several characteristics simulta neously. The solution currently suggested for multiple characteristic se lection is index selection, the other possibilities being tandem selection and independent culling. Building up a successful selection index de mands economic weighting of the characteristics included in the index, and reliable estimates of the heritabilities and genetic correlations be tween the characteristics (Cotterill and Dean 1990). The selection for two characteristics with a strong unfavourable genetic correlation cannot be effective for the both characteristics at the same time. Thus including a new trait in the selection programme demands, not only an investigation of the additive genetic variation of the characteristic itself, but also a comprehensive study of the relationships with the characteristics that have already been included. Wood quality characteristics are often not manifested in young seed ings. In the case of such characteristics, investigation of the age-age type genetic correlations is not as relevant as with the early selection of pro duction characteristics. However, in order to keep the interval between breeding generations as short as possible, there is a need for early selec tion also in the quality characteristics. A successful early indirect selec tion requires the breeder to find, either by carefully investigating the ge netics and physiological background of the target characteristic or merely by screening several characteristics, a high genetic correlation between the target and any other characteristic which is manifested at an early age of the tree with a high heritability. In the case of heartwood properties, the question about early indirect selection is not only important, but also exciting, because the wood tissues of the future heartwood already exist in fairly young stems, but still possess the properties of sapwood (Fig. 1)! 1.4 The current criteria of Scots pine wood quality The difficulty of determining wood quality has sometimes been overem phasised. Nevertheless, 'high quality' wood is a fairly simple concept: it means a good suitability for the purpose the wood is to be used for (Zobel and van Buijtenen 1989, Kellomäki et ai. 1992) or, in economic terms, it is something the customers are willing to pay extra for (Venä läinen et ai. 1996). However, an important consequence of the definition 14 Decay resistance of Scots pine heartwood is that different, sometimes even contradictory, properties are deemed to mean high quality, depending on the purpose the wood is used for. The classical example of this are the studies on the chemical properties of Scots pine heartwood which led to the discovering of the phenolic extrac tives playing a role in decay resistance. The studies were initiated because heartwood was found to be detrimental for the sulphite pulping process in the paper industry (Erdtman 193%). Since then, Nordic pulp engineers have turned their main attention to the fibres of other tree species, such as spruce, birch and aspen (Paavilainen 2002), and thus the mechanical in dustry, together with its customers, has alone set the quality requirements for Scots pine wood. The Nordic grading rules for sawn timber, dating from the 1930's and updated in the 1980's, set upper limits first of all for the size and number of knots (Pohjoismainen ... 1994). The occurrence of actual defects (such as scars and reaction wood) is also limited, but there is no strict upper limit for the growth rate as such. In practice, the high phenotypic correla tion between rapid growth and thick branches has led to a rule of thumb: the inner knot quality of a log is predicted by the width of the annual rings visible at the end of the log (Heiskanen 1965, Halinen 1985). Rapid growth as such, leading to wide annual rings, low late wood proportion and low basic density, may decrease the quality of timber only in those usages where high strength is required for beams. Overall, it can be in ferred that rapid growth and high timber quality are not logically contra dictory properties to be included in the same ideotype. Neither does there seem to be any evidence that a Scots pine tree, growing rapidly and pro ducing wood which fulfils the current quality criteria, would be a biologi cal impossibility. 1.5 Natural durability 1.5.1 The 'rehabilitation' of natural durability In the early days, before the invention of artificial wood preservatives, the people who used wood were fully aware of the differences in the natural durability between tree species and between different parts of the stem, as well as the use of natural wood extractives as preservatives (Plinius -A.D.77, Richardson 1978, Hillis 1987). Nowadays, when the reputation of coal-tar distillate 'creosote', patented in 1838, and the range of metallic salt preservatives, such as copper-chromium-arsenic (CCA) products, that have been developed during the 19 th and 20 th centuries (Richardson 1978, Zabel and Morrell 1992), have been questioned, there has been a recur rence of interest in the natural durability of wood. In Finland the rules Martti Venäläinen 15 concerning the production, sale, use (decisions of Finnish Environment Institute, Suomen ympäristökeskus) and especially the disposal of im pregnated wood were considerably tightened in 2002 (Suomen säädös kokoelma 1128/2001, 1129/2001). On the other hand, consumers have begun to appreciate the advantages of ecologically benign over forever lasting wood, for example in garden fittings and playground structures (Life Cycle Assessment). In massive buildings such as houses, a long ser vice life for untreated wooden elements can best be guaranteed by keep ing the wood relatively dry, i.e. by maintaining the moisture content be low 20 % (Viitanen 1996). In general, it will be difficult to avoid the use of impregnated wood in places with the highest decay risk (e.g. railway sleepers and electricity transmission poles in contact with the soil), as well as in constructions that are difficult and therefore expensive to re pair, or the weakening of which may cause a safety hazard (e.g. guard rails of a bridge) (Richardson 1978). The variation in the durability of im pregnated wood is smaller than that of natural durability, which increases the value of impregnated wood as a standardised building material. Among the tree species growing in Finland the trees considered to have wood of the highest natural durability are yew (Taxus baccata L.) and oak (Quercus robur L.) (Adopted European ... 1994b). However, they have no use on a practical scale because the wood production of them in our climate is relatively limited. The medium level of durability is represented by the heartwood of larch (Larix sp) and Scots pine. Ac cording to the five-class scale of the European standard EN 350-2, Scots pine heartwood is classified as 3-4, i.e. moderately to slightly durable against wood-destroying fungi, while Siberian larch (L. sibirica Ledeb.) is not included in the list of species (Adopted European ... 1994b). Global trade makes it possible to utilise non-domestic durable softwoods, such as western red cedar (Thuja plicata Donn) from North America, or the tropical, very durable hardwoods, in the Nordic countries. However, the easy accessibility of raw material favours the use and price of domes tic species. Moreover, if the wood, even though less durable, is produced near to the site where it will be used, it gains some additional value as re gards ecological indicators (Material Flow Analysis). 1.5.2 Factors affecting natural durability The wide variation in the durability among and within species has led to a number of scientific reports, starting in the 1920's (see the early refer ences given by Zabel and Morrel, 1992), as well as speculation about the factors that may play role in natural durability. These factors are mainly associated with the wood extractives, "the principal source of decay resis- 16 Decay resistance of Scots pine heartwood tance", that inhibit the primary metabolism or degradation processes of the fungi, or with the permeability of the wood for water, air and fungal hyphae (Scheffer and Cowling 1966). Approximately the same factors are involved in the formation of heartwood (Rudman 1966). Other factors, which affect the suitability of wood as a living environment for fungi and the variation of which could thus contribute to the durability variation, are lignification of the wood cell walls, degree of crystallinity of the cellu lose, nitrogen content of the wood, and the depletion of reserve food ma terials (Scheffer and Cowling 1966, Gref et al. 2000). The wood degradation mechanisms of brown-rot fungi have been gradually elucidated during the past two decades (Highley and Dashek 1998). Both oxidative free radicals (Ritschkoff 1996) and cellulolytic en zymes (Eriksson et al. 1990, Maijala 2000) are involved in the processes, as soon as there is sufficient water within the wood cells for the function ing of fungal hyphae (Zabel and Morrell 1992, Viitanen 1996). The un derstanding of the oxidative degradation mechanisms especially has re vealed new potential sources of durability variation, such as the function ing of iron ions (Fekete et al. 1989, Goodell et al. 1997, Hyde and Wood 1997, Goodell and Jellison 1998, Xu and Goodell 2001), or the role of extractives as antioxidants (Schultz et al. 1997, Schultz and Nicholas 2000). According to Zabel and Morrell (1992), there are four major groups of heartwood extractives that include compounds known or believed to con tribute to the natural durability: phenolics including stilbenes and flavon oids, tannins, terpenoids and tropolones. A more comprehensive grouping of heartwood extractives is presented by Hillis (1987), and the biosynthe sis and biological activity of them by Seigler (1998). In the case of Scots pine the relevant groups of extractives are stilbenes and terpenoids. The compound called pinosylvin (PS), named after Pinus sylvestris by Erdman (1939 a), and its derivates pinosylvin monomethylether (PSM) and pinosylvin dimethylether (PSD), represent stilbenes and are typical for the heartwood of Pinus species. Very low concentrations of free forms of PS and PSM have shown to be toxic for fungi (Rennerfelt 1943, 1945). Based on his investigation on the metabolism of microorganisms, Lyr (1961) concluded that PSM acts as an uncoupling toxin that inhibits the oxidative phosphorylation, and assumed that PS acts in an equal manner. However, the role of stilbenes in situ, as integrated in the wood substrate, has been a puzzling and even argued conundrum (Erdtman and Rennerfelt 1944, Rennerfelt 1947, Rudman 1963, Loman 1970a,b, Hart and Shrimp ton 1979, Hart 1981, Celimene et al. 1999, Schultz et al. 1997, Schultz and Nicholas 2000). The oleoresin of Scots pine contains two types of terpenoids: volatile monoterpenoids, such as a-pinene and 3-carene, and non-volatile diterpe- Martti Venäläinen 17 noids, resin acids (Sjöström 1993). The fungitoxic monoterpenoids in oleoresin makes it important for the active defence of pines (Flodin and Fries 1978, Yamada 1992 and references therein). It has also been sug gested that the resin acids, that remain in timber during the drying proc esses, play a role in decay resistance (Hart et al. 1975 and references therein). As regards the mechanism of inhibition, the resin acids are sup posed to make the wood hydrophobic (Verrall 1938, Eberhardt et al. 1994) or act as weak fungicides (Hartman et al. 1981, Micales et al. 1994). In any case, the decay resistance of sawn timber can be based only on passive defence mechanisms (see the definition in Merrill 1992) pos sessed by tissues that have no living cells. Induced processes that are im portant for the active defence of the living tissues of standing trees can no longer occur. 1.5.3 In vitro methods for studying natural durability The accelerated testing of preservatives against wood destroying fungi has a long tradition, and several testing procedures have been developed (see Zabel and Morrell 1992). In today's Europe the testing of wood pre servatives against Basidiomycetes has been standardised by the norm EN 113 (European standard EN 113, 1996). The accelerated testing procedure for the natural durability of wood, based on EN 113, has been standard ised by the norm EN 350-1 (Adopted European ... 1994 a). In the stan dardised test, the size of the wood block is 50 x 25 x 15 mm and the blocks are incubated for 16 weeks in contact with pure cultures of three different fungus species (which in the case of softwood are Gloeophyllwn trabeum (Persoon ex Fries) Murrill, Serpula lacrymans (Schumacher ex Fries) S.F Gray, Poria placenta (Fries) Cooke sensu J.Eriksson). Accord ing to Hannu Viitanen from VTT (pers. comm. 11.1.2002), the use of Co niophora puteana (Schumacher ex Fries) Karsten, which is an obligatory brown-rot species in the EN 113 test, is today also accepted in the testing of natural durability. Reference specimens from Pinus sylvestris sapwood are obligatory when softwoods are tested. The mass loss of the samples is used as an inverse measure of the decay resistance. The results are ex pressed as the ratio of the loss in mass and the oven dry (103° C) initial mass of the sample, i.e. as mass loss percentage (%). The first obstacle in using the standardised testing method in estimat ing the genetic parameters of decay resistance, and in testing breeding material, is the destructive sampling. In the breeding procedures the trees are needed, after the testing, to produce the next generation (i.e. in pollen and seed production) or to be used in vegetative propagation. Thus they 18 Decay resistance of Scots pine heartwood have to remain alive and vigorous for several years after sampling. The other disadvantage of the standardised method are the experimental costs. The number of samples needed to meet the requirements of statistical precision in parameter estimation is high, several hundred at the mini mum. Thus testing would occupy a lot of laboratory space for a long pe riod, especially if parallel samples are tested against several fungus spe cies. 1.6 Aim of the study The aim of this doctoral thesis study was to evaluate the biological possi bility of improving the decay resistance of Scots pine heartwood timber by means of forest tree breeding. In order to be able to draw conclusions on this topic, the following questions had to be addressed. The Roman numerals refer to the original articles, including empirical results, on which the answers and discussion were based on. Fundamental questions: a) Is it possible to assess the decay resistance of the timber, i.e. sawn and dried wood product, of a standing tree ? (I, 11, 111, VI) b) What is the magnitude of the natural phenotypic variation in the decay resistance of Scots pine heartwood timber? (11, 111, VI) c) How large is the additive genetic component of variance? (11, III) Further questions provided additive genetic variation exists: d) In the implementation of the Scots pine breeding programme, is it pos sible to select directly for the target characteristic, i.e. the decay resis tance of timber? (11, 111, VI) e) Which characteristics can be used in the indirect selection and in the early indirect selection ? (IV, V, VI) f) What are the genetic correlations between decay resistance and those used in the indirect selection? What are the genetic correlations between decay resistance and those already included in the Finnish Scots pine breeding programme? (discussion based mainly on the literature) g) What is the role of the environment in the recurrent selection of Scots pine breeding material and in the management of the stands established with the improved regeneration material. (11, III) Martti Venäläinen 19 2. Material and methods 2.1 Wood material The planning of sampling in studies such as the present one is trouble some if the organisation of different types of woody tissue within a stand ing tree is not fully understood. Figures 3 and 4 in the Appendix provide information about the age of the tissues and the internal variation of the stem. The sampling in young trees especially is fundamentally associated with the concept of juvenile wood (see e.g. Zobel and van Buijtenen 1989, Thörnquist 1993). The most accepted explanation for the formation of juvenile wood is the juvenile stage of the cambium; ageing of the cam bium gradually leads to the formation of mature wood. The main differ ences between juvenile and mature wood lie in the dimensions of the tra cheids, the basic density and the microfibril angle. The duration of the ju venile stage in Scots pine is poorly known, but it can be assumed to con tinue for 20-30 years (Thörnquist 1993). Thus the core of the stem, i.e. approximately the first 20-30 annual rings around the pith, at each rela tive height of the tree consists of juvenile wood. Thus, when the forma tion of heartwood begins, the first 20-30 annual rings around the pith will be juvenile heartwood. The material used to develop the accelerated decay test for standing trees was obtained from a Siberian larch clonal seed orchard at Jäm sänkoski (I). The samples were taken with a 5 mm diameter, increment core borer. A 40 mm-long piece of the outer heartwood section of the core was separated for the decay test. In studies II and 111 an equal-sized increment core borer was used to take samples from standing Scots pines. In the planning of the sampling, the main emphasis was put on the precise estimation of the genetic parameters. The progeny tests established with half-sib families are well suited for this purpose. However, a compromise was needed in order to adjust the number of families, and trees per family to be sampled (Xie and Mosjidis 1997), as well as the number of samples per tree, with the total number of samples that could be included in the decay test at same time (about 1000). On the other hand, in order to ob tain as mature wood as possible the stand had to be selected from among the oldest progeny tests, the designs of which set their own limitations on the sampling. At Kerimäki, 10 trees from 25 families (totalling 250 trees) were sampled from a 32-year-old progeny test (II), and at Korpilahti 413 trees from a total of 26 families in a 34-year-old progeny test were in cluded in the statistical analysis (III). In study 11, four sapwood samples and one heartwood sample per tree were studied. In study 111, only one heartwood sample per tree was subjected to the decay fungus. For estima- 20 Decay resistance of Scots pine heartwood tion of the offspring-parent regression, 20 clones representing the mother trees of the half-sib families of the Korpilahti progeny test were sampled in the clonal archive in Punkaharju (III). In order to study the hypotheses concerning the relationships between decay resistance and other wood characteristics, derived from the frame work presented in Figure 2 (Chapter 3.6.2), a new set of sample trees was selected from among the trees included in studies II and 111. Two groups of trees with extreme rates of juvenile heartwood decay were obtained by culling the tails of the in vitro mass-loss distributions. The group of trees with slowly decaying heartwood, referred to later on as resistant, com prised 10 trees from Kerimäki and 10 trees from Korpilahti. The 20-tree group with rapidly decaying heartwood, referred to later on as suscepti ble, were also derived from the same locations (Table 2 in IV). Sampling was carried out on an individual tree basis, but only one tree from a single half-sib family was accepted in order to avoid a kin structure among the sampled trees. This time the sampling was carried out with an 8 mm in crement core borer as described in study IV. The sampling procedure limited the use of parametric statistical meth ods in the data analysis in studies IV and V. In contrast, the sampling for the study VI was random as regards the decay resistance and the distribu tion of the samples was continuous, thus allowing better use of parametric statistical methods. Two additional elements were included in the mate rial for study VI: variation in an old natural stand, and the radial variation in large-sized mature trees. Destructive sampling was applied by taking disks from twenty felled trees and by re-sampling the disks (figure 1 in VI) 2.2 Decay tests In order to avoid the problems associated with the use of standardised de cay test EN 113 in studies like the present one, VTT Building and trans port developed a simplified testing method. The main advantage of this malt agar plate method, described in detail in article I, is the possibility to study standing trees. This was enabled by the use of an increment core borer in the sampling. The small size of the specimen permitted a short incubation time, 6-8 weeks, and only one brown rot fungus, C. puteana (strain BAM Ebw. 15), was used instead of the three fungus species. The 5 mm x 35 - 40 mm increment core section was used in studies II and 111. The destructive sampling in study VI made it possible to use a small (5x15x30 mm) block as a specimen, but otherwise the decay testing pro cedure was similar in studies I, 11, 111 and VI. Martti Venäläinen 21 According to the norm EN 113, the result of the decay test is ex pressed as the relative mass loss (loss in mass/initial dry mass x 100 %). However, because of the large natural variation of basic density, the rela tive mass loss alone does not provide sufficient information in the dura bility analysis. In the present study the mass loss was expressed as an ab solute measure per fresh wood volume (mg/cm 3 ), which better describes the degrading activity of the fungus. Only in study 111 the genetic pa rameters were also estimated for the relative mass loss. Irrespective of whether the variable is either absolute or relative, the information about the basic density of the samples should always be included in the report ing. 2.3 Quantification of extractives For the quantification of resin acids (IV, VI), wood powder was extracted with petroleum ether-diethyl ether following the procedures of Gref and Ericsson (1985). The extracts were analysed by gas chromatography - mass spectrometry (GC-MS) as described by Manninen et ai. (2002). For quanti fication of individual resin acids, calibrations were made using known amounts of pure resin acids and the response factors were determined for each substance relative to known amounts of the internal standard (hepta decanoic acid). For the analysis of the total concentration of all phenolic secondary compounds, wood powder was extracted with 80 % (v/v) acetone. The phenolics were determined by the Folin-Ciocalteu technique using tannic acid as a standard (Julkunen-Tiitto 1985, Turtola et ai. 2002). The con centrations of soluble phenolics other than the stilbenes PS and PSM are low in Scots pine heartwood, and thus the result of the Folin-Ciocalteu determination is assumed to mainly depict the concentration of stilbenes. For the quantification of the individual stilbenes, pinosylvin (PS), pino sylvin monomethyl ether (PSM), and pinosylvin dimethyl ether (PSD), two different techniques were used. In study V wood powder was extracted with acetone in a mini-Soxhlet apparatus for 6 hrs. An internal standard, diethylstilbestrol, was added and the extracts were evaporated to dryness and stored under nitrogen. Trimethylsilyl esters were prepared, after which PS, PSM, and PSD were determined by gas liquid chromatography - mass spectrometry (GLC-MS). In study VI, wood powder was extracted with 80 % (v/v) methanol. The extraction was carried out in tubes with mixing for 30 minutes. Vanillin was used as an internal standard. Samples were centrifuged and the supernatants were analysed by high performance liquid cromatography (HPLC). The analysis was performed as described by Lieutier et al. (1996). Peak areas were used to quantify the individual 22 Decay resistance of Scots pine heartwood substances PS and PSM, and the results (mg/g dry wt) were calculated relative to known amounts of internal standard. However, in study VI, the final results of all the chemical analyses were expressed, correspondingly to the absolute mass loss, as concentration per fresh volume of wood. 2.4 Water sorption capacity Water bound to hygroscopic cell wall constituents and into voids of wood of radius less than 1.5 pm is called adsorbed water. This critical point of sorption is called the fibre saturation point. It represents a water potential of-0.1 MPa and, in theory, a relative humidity of 99.93 % (Griffin 1977). The water present in the cell lumens and intercellular space is called free or absorbed water (Walker 1993). The adsorption capacity of the Scots pine heartwood was first (IV) stud ied using increment core samples taken from the Korpilahti progeny test in January 2001 from the same 10 resistant and 10 susceptible trees on which the analysis of extractives had been performed. After drying at 60° C for 48 hours and weighing, the samples were conditioned at a relative humidity (RH) of 100 % at 24° C. After two weeks of conditioning the samples were weighed, oven dried at 103° C for 24 hours, and reweighed. The quantity of water in the equilibrium stage was expressed as mg/cm 1 , and the equilibrium moisture content, MC %, as the mass of water at RH 100 %, divided by the oven dry mass of wood at 103° C. The moisture content can be regarded as the quantity of water standardised by the wood density. The water sorption capacity was then studied in two different kinds of experiment using 5 x 15 x 30 mm-sized, wood blocks (VI). The adsorp tion was first determined in a tightly closed steel tank that was half-filled with tap water (+25° C). The wood blocks (dried at 60° C for 48 hours) were placed on steel racks immediately above the water surface. The blocks were weighed at increasing intervals from 4 hours up until 14 days after the start of the test, and then dried at 103° C to obtain the dry mass. The results were expressed as the moisture content values, MC %. The same blocks were then immersed in water. The mass of the wet blocks was measured at increasing intervals from 1 hour to 7 days after the start of the test. The quantity of sorbed water was expressed as the gross mass of water (i.e. adsorbed and absorbed) per fresh volume of wood (mg/cm 3 ). 2.5 Statistical analysis In the studies II and 111, in which variance components were estimated the models of analysis were based on individual tree observations, assum- Martti Venäläinen 23 ing that all effects associated with the observations were incorrelated. In contiguous plot designs, there is the possibility of spatial auto-correlation. If several trees per plot are sampled and the assumption of independence is wrong, the additive genetic component of variance is overestimated. In the case of wood characteristics the danger of this bias was not consid ered to be serious. The measurable phenotypic value (P) an individual tree is assumed to be a sum of an additive genetic effect (A), and an environmental effect (E) which here also includes the possible non-additive genetic effects (P=A+E). Correspondingly, the observed phenotypic variance splits into additive genetic and environmental components: aP 2 = cr/+ = + o~E . The narrow-sense heritability is calculated as h 2 =G^/Gp Variance components were estimated with the REML technique utilizing the MIXED procedure of the SAS system (SAS Institute Inc. 1992). The random model equation was where ybjj is an observation for a tree; nis a fixed general mean; ab and bf are the random effects of the block b and the family /, which vary from 1 to 5 and from 1 to 25, respectively; cbf is the random block x family interaction effect; and ebfi is the residual random ef fect of tree i in plot bf. All the effects are assumed to be independ ent. The random effects were assumed to be normally distributed with the expectation zero and the variance g£, oj, ojy, and cL [N(O, gI), etc.]. The individual-tree heritability for half-sib proge nies was estimated as This estimate of the heritability is used when no corrections for block effects are applied prior to selection (Cotterill 1987), and when the purpose is to describe the effects of the sources of varia tion in a particular experiment. The MIXED procedure of the SAS system also provided approxi mate standard deviations for the estimated variance components. The standard deviation for the heritability estimate was calculated using "Dickerson's approximation" (e.g.. Dieters et al. 1995), and for the additive genetic component Gd: =4ad:. The coefficient of additive genetic variation was calculated by dividing oA by the overall mean value of the trait. Results There was minor genetic variation in the durability of Scots pine sapwood against a single strain of the brown rot fungus C. puteana. The heritability for the sapwood mass loss and CV A were among the lowest in the range of traits studied (Table 1). The heritability estimates for the heart wood mass loss were of the same low order of magnitude as for the sapwood. However, for the heartwood mass loss was high, 10.6% (Table 1). The mean mass loss differed considerably between the sapwood and heartwood, the heartwood losing less mass than the sapwood (Table 1). The difference between the av erage mass losses was 33.6 mg-cm" 3 during the 6-week de cay test. The range for the heartwood mass loss was 198.8 mg-cm" 3 compared with 72.4 mg-cm" 3 for sapwood (Fig. 1). There was unimportant phenotypic correlation be tween the sapwood and heartwood mass loss at both the in dividual tree (r = -0.062) and the family mean (r = 0.047, SD = 0.507) level. Because of the low number of families in the study and low heritability of the sapwood and heartwood mass loss, it was not reasonable to estimate the genetic cor relation between these parameters. At least in this progeny test, indirect selection for the heartwood mass loss on the basis of sapwood mass loss would not be effective. A number of wood and growth traits, including the distance to the crown limit, had heritabilities ranging from moderate to high. However, the stem diameter at breast height, sapwood width, and the heartwood radius, had low heritabilities (Ta ble 1). Among the wood traits, the number of annual rings in sapwood and heartwood had heritabilities of 0.47 and 0.39, respectively, and the number of heartwood annual rings also had a high CV A (Table 1). The density of the heartwood reached the highest heritability among all traits (h 2 = 0.57). [l] ybfi - + +ba bf + cbf + ebJi [2] h 2 =- + 0/ + Öi/ + ö; Harju et ai. 1247 © 2001 NRC Canada Table 1. Estimates of genetic parameters of wood and growth traits. Note: Sample size (AO and mean values for the traits studied and the estimates for the phenotypic and additive genetic variances (o>, for the narrow-sense heritability (h 2 = h\/h 2 P), and the additive genetic component of variation (CVA , xlOO). "Values are estimates ± SD; the p value for the approximate test about whether the additive component of variance is zero is given in parentheses. ''Values are estimates ± SD. 'Latewood percentage is for 10 outermost annual rings. Discussion We obtained low narrow-sense heritabilities for mass loss in both sapwood and heartwood, which means that, in this progeny test, the effects of environmental factors on the mass loss were much more important than the genetic deter mination. Genetic variation of these traits may actually be low or some of the environmental factors, which effected the wood durability, were unexpectedly heterogeneous in this progeny test. Nuisance factors connected to the laboratory experiment might have been one reason for the high propor tion of unexplained residual variation. A suitable covariate describing the fungal activity in each agar plate would have probably decreased the unexplained variation. However, a similar laboratory experiment with Siberian larch gave a broad sense heritability of 0.39 (Venäläinen et ai. 2001), and there fore, we do not expect the experimental inaccuracy to be the main reason for low heritabilities in the mass loss. Both the phenotypic variation and CVA for sapwood mass loss were low, indicating minor variation in the passive de fence system of sapwood against fungal attack. In the sap wood, the defence mechanisms involved in the wounding of a standing tree are based on active processes, i.e., when the sapwood is wounded the injured area will be compartmental ized by changes in wood anatomy and chemistry. The poor passive resistance of the sapwood does not necessarily indi cate poor active or induced resistance of the standing trees against plant pathogens. Our results revealed considerable phenotypic variation and a high CVA in the decay resistance of heartwood. A large fraction of the cores did not lose mass at all. As a similar distribution was not found for the sapwood samples, the cause of this is most probably due to the wood material rather than the experimental conditions during the decay test. In western redcedar (Thuja plicata Donn.), the decay resistance of the wood near the pith was extremely variable (Deßell et al. 1999). In western redcedar the decay resis tance has been connected to the presence of toxic extract ives, tropolones. In the study of Deßell et al. (1999), the highest and the most variable mass losses were associated with samples containing very low tropolone levels, while the samples with more tropolones had a lower, less variable de gree of mass loss. Eberhardt et al. (1994) has reported that resin acids extracted from Pinus ponderosa Dougl. ex Laws, and Pinus banksiana Lamb, seed cones provided decay re sistance against white rot fungi as a result of their hydropho bic properties rather than their general toxicity when applied to test blocks of sweetgum (Liquidambar styraciflua L.) wood. In Scots pine, a phenolic compound, pinosylvin. has been shown to cause the decay resistance of heartwood (e.g., Rennerfelt 1956; Rennerfelt and Nacht 1955). However. Hart and Shrimpton (1979) emphasize caution in such inter pretations because of the complicated nature of the in situ interaction between decay fungi and stilbenes to which pino sylvin belongs. However, the reason for the great phenotypic variation in heartwood mass loss in our material remains unclear for the present. One reason may be that the sample included both straight and curved trees, which may have dif ferent anatomical structure and chemical composition and thus a varying effect on fungal activity. It also could be hy pothesized that young Scots pine trees in the early stages of heartwood formation differ with respect to the timing of events connected to decay resistance. The relatively high CV A , accompanied by a low heritability for heartwood mass loss, necessitates selection with intensive testing of the geno types. We did not estimate genetic correlation between sapwood and heartwood mass loss, because the family components of the variance were low, with standard deviations being three to seven-fold larger than the component itself, respectively. Unimportant phenotypic correlation was found at the indi vidual and the family mean level. Thus, at least outside the growing season, it is not possible to perform indirect selec tion on the basis of measurements of sapwood mass loss. Trait N Mean a? G a" h- b CV„ % Sapwood Mass loss, mg-cm" 3 248 113.6 144.79 4.74± 16.00 (0.77) 0.03±0. 11 1.9 Wood density, mg-cm" 3 248 391.0 607.37 204.78± 140.78 (0.14) 0.34 ±0.23 3.7 No. of annual rings 238 17.4 1.79 0.85±0.49 (0.08) 0.47±0.27 5.3 Width, cm 245 5.3 0.61 0.03±0. 10 (0.75) 0.05±0. 16 3.3 Latewood, %' 246 39.2 22.26 12.61 ±4.43 (0.19) 0.26±0.20 6.1 Heartwood Mass loss, mg-cm" 3 248 80.0 3677.03 71.87±534.36 (0.89) 0.02±0.14 10.6 Wood density, mg-cm" 3 248 337.1 1014.98 580.90±286.29 (0.04) 0.57±0.28 7.1 No. of annual rings 241 6.4 1.70 0.66±0.43 (0.13) 0.39±0.25 12.6 Radius, cm 242 3.0 0.61 0.04±0.10 (0.72) 0.06±0. 16 6.2 Latewood, % 242 13.2 6.10 2.7 3± 1.47 (0.06) 0.45±0.24 5.4 Height, m 249 15.3 1.58 0.38±0.33 (0.25) 0.24±0.21 4.0 Crown limit, m 249 7.6 1.54 0.58±0.37 (0.11) 0.38±0.24 10. 1 Diameter (1.3 m), cm 249 18.5 4.54 0.00±0.00 (—) 0.00±0.00 0.0 1248 Can. J. For. Res. Vol. 31, 2001 © 2001 NRC Canada Heartwood formation is evidently an age-dependent pro cess. Lappi-Seppälä (1952) studied naturally regenerated, 50- to 150-year-old Scots pines growing on relatively infertile sites. He concluded that heartwood formation had started at the age of 30-40 years. However, Fries and Ericsson (1998) found heartwood diameters ranging between 5.5 and 53.5 mm at a height of 80 cm above the ground in 25-year-old sap lings in a progeny test. Moreover, the study of Björklund (1999) suggests that heartwood formation begins in the pith when the cambium is about 15 years old. In our material, the average numbers of annual rings in heartwood and sapwood at breast height were 6 and 17, respectively. Heartwood di ameter has been reported to have a high heritability (Fries and Ericsson 1998; Ericsson and Fries 1999). In our study, the somewhat curved growth habit of the trees may have caused the low narrow-sense heritability for the heartwood radius. In practice, the highest heritabilities are obtained in prog eny tests where the environment is homogeneous, and to a great extent the individual differences are due to additive ge netic effects. We suggest that, in this study, one environmen tally induced factor which probably affected the mass loss through anatomical structure and chemical composition of the wood, was the fact that all the sampled boles were not straight. There were marked differences in the mean width of the annual rings on different sides of the stem, indicating the existence of reaction wood. It has also been reported that compression wood is more common in juvenile wood than in mature wood (e.g., Dadswell 1958). Thus, it is apparent that the wood samples used in our study contained reaction wood. Blanchette et al. (1994) reported that, compared with nor mal wood, the compression wood of Abies balsamea (L.) Mill., Picea mariana (Mill.) BSP, and Pinus strobus L. was more resistant to decay caused by white and brown rot fungi, but the reason for this remained unclear. It has also been reported that compression wood contains higher con centrations of lignin than normal wood (Panshin and de Zeeuw 1980, p. 308, and references therein; Zobel and van Buijtenen 1989, p. 145, and references therein; Blanchette et al. 1994). On the other hand, the opposite wood may have higher cellulose and lower lignin content than normal wood (Zobel and van Buijtenen 1989, p. 144, and references therein; Haygreen and Bowyer 1996, p. 112). Thus, the curved growth form of the stems in our study may have re sulted in changed lignin and cellulose concentrations com pared with those of normal wood, thus affecting the amount of substrate available for cellulose decomposing, brown rot fungus. Moreover, our sample was most probably a mixture of increment cores from normal, compression, and opposite wood. As a result of the uncontrolled and irregular occur rence of the different types of the wood, the residual vari ance for heartwood mass loss became very high compared with the additive genetic variance, resulting in low narrow sense heritability. In our study we did not find any connection between the visual observations of the stem form and decay resistance against C. puteana. However, because we did not determine the presence of reaction wood directly from the wood samples, the relationship between decay and reaction wood could not be studied. Decay resistance as such is a complicated combination of traits that are poorly known. The traits and the progeny test we studied did not promise good possibilities for increasing the decay resistance of Scots pine wood against brown rot fungus through tree improvement. Caution is needed when generalizing about genetic parameters estimated from a sin gle progeny test (see e.g., Haapanen et ai. 1997; Jacquard 1983). Acknowledgements The progeny test was planned by the late Risto Sarvas and the late Reino Saarnio. Heikki Paajanen, Heikki Kinnunen, and Teemu Moilanen took the increment cores. Outi Kylliäinen and Markku Tiainen carried out the microscopic study at Joensuu Research Station. The decay resistance tests were performed by Liisa Seppänen at VTT Building and Transport. Eija Matikainen assisted with the data files, and John Derome re vised the language. We thank all the above-mentioned collabo rators. We are grateful to Tore Ericsson and Matti Haapanen for discussions concerning the statistical models and their ap plications. We thank Seppo Ruotsalainen for reading the manu script with care and making many constructive comments. Two anonymous referees made valuable comments on the manu script. The Academy of Finland (Resource Council for the Environment and Natural Resources, project No. 43140) supported this study, which belongs to the Finnish Forest Cluster Research Programme Wood Wisdom (1998-2001). References Björklund, L. 1999. Identifying heartwood-rich stands or stems of Pinus sylvestris by using inventory data. Silva Fenn. 33(2): 119-129. Björklund, L., and Walfridsson, E. 1993. Properties of Scots pine wood in Sweden—basic density, heartwood, moisture and bark content. [ln Swedish with English summary.] Department of Forest Products, Swedish University of Agricultural Sciences, Uppsala. Rep. 234. Blanchette, R.A., Obst, J.R., and Timell, T.E. 1994. Biodegradation of compression wood and tension wood by white and brown rot fungi. Holzforschung, 48: 34-42. Bruce, A. 1998. Biological control of wood decay. In Forest prod ucts biotechnology. Edited by A. Bruce and J.W. Palfreyman. Taylor & Francis Ltd., London, pp. 251-266. Climent, J., Gil, L„ and Pardos, J. 1993. Heartwood and sapwood de velopment and its relationship to growth and environment in Pinus canadensis Chr. Sm ex DC. For. Ecol. Manage. 59: 165-174. Cotterill, P.P. 1987. On estimating heritability according to practi cal applications. Silvae Genet. 36: 46-48. Dadswell, H.E. 1958. Wood structure variations occuring during tree growth and their influence, on properties. J. Inst. Wood Sci. 1: 11-24. Deßell, J.D., Morrell, J.J., and Gartner, B.L. 1999. Within-stem variation in tropolone content and decay resistance of second growth western redcedar. For. Sci. 45(2): 101-107. Dieters. M.J., White, T.L., Littell. R.C., and Hodge, G.R. 1995. Application of approximate variances of variance components and their ratios in genetic tests. Theor. Appl. Genet. 91: 15-24. Eberhardt, T.L., Han, J.S., Micales, J.A., and Young, R.A. 1994. Decay resistance in conifer seed cones: role of resin acids as in hibitors of decomposition by white-rot fungi. Holzforschung. 48: 278-284. 1249 Harju et ai O 2001 NRC Canada Ericsson, T., and Fries, A. 1999. High heritability for heartvvood in north Swedish Scots pine. Theor. Appi. Genet. 98: 732-735. Fries, A., and Ericsson, T. 1998. Genetic parameters in diallel crossed Scots pine favour heartwood formation breeding objec tives. Can. J. For. Res. 28: 937-941. Haapanen, M., Veiling, P., and Annala, M.-L. 1997. Progeny trial estimates of genetic parameters for growth and quality traits in Scots pine. Silva Fenn. 31(1): 3-12. Hart, J.H.. and Shrimpton, D.M. 1979. Role of stilbenes in resis tance of wood decay. Phytopathology. 69: 1138-1143. Haygreen, J.G., and Bowyer, J.L. 1996. Forest products and wood science: an introduction. 3rd ed. lowa University Press, Ames. Hillis, W.E., and Ditchburne, N. 1974. The prediction of heartwood diameter in radiata pine trees. Can. J. For. Res. 4: 524-529. Jacquard. A. 1983. Heritability: one word, three concepts. Biometrics. 39: 465-477. Kellomäki, S. 1981. Quality of pine logs and proportion of heart wood as related to properties of the logs. [ln Finnish with Eng lish summary.] Folia For. 489. pp. 1-13. Kramer, P.J., and Kozlowski, T.T. 1979. Physiology of woody plants. Academic Press. Inc.. Orlando, Fla. Kärkkäinen, M. 1972. On the proportion of heartwood in Norway spruce (Picea abies (L.) Karst.) and Scots pine (Pinus sylvestris L.). [ln Finnish with English summary.] Silva Fenn. 6(3): 193-208. Kärkkäinen, M. 1976. Wood science prerequisites for the weight measurement of pine and spruce logs. [ln Finnish with English summary.] Commun. Inst. For. Fenn. 89(1). pp. 1-58. Lappi-Seppälä, M. 1952. Über Verkernung und Stammform der Kiefer. [ln Finnish with German summary.] Comm. Inst. For. Fenn. 40(25). pp. 1-26. Löyttyniemi, K. 1986. On natural durability of pine heartwood. lln Finnish with English summary.] Metsäteknologian tutkimusosasto, Puuntutkimussuunta, Finland. Metsäntutkimuslaitoksen tiedonan toja 231. Panshin, A.J., and de Zeeuw, C. 1980. Textbook of wood technol ogy. Structure, identification, properties, and uses of the commer cial woods of the United States and Canada. 4th ed. McGraw- Hill. Inc., New York. Rennerfelt. E. 1956. The natural resistance to decay of certain co nifers. Friesia, 3(5): 361-365. Rennerfelt, E., and Nacht, G. 1955. The fungicidal activity of some constituents from heartwood of conifers. Sven. Bot. Tidskr. 49: 419-432. SAS Institute Inc. 1992. SAS° technical report P-229, SAS/STAT® software: changes and enhancements, release 6.07 ed. SAS In stitute Inc., Cary, N.C. Schmidtling, R.C., and Amburgey, T.L. 1982. Genetic variation in decay susceptibility and its relationship to growth and specific gravity in loblolly pine. Holzforschung. 36(3): 159-161. Tamminen. Z. 1962. Moisture content, density and other wood properties of wood and bark of Scots pine. [ln Swedish with English summary.] Department of Forest Products, Royal Col lege of Forestry, Stockholm, Sweden. Rep. 41. Uusvaara, O. 1974. Wood quality in plantation-grown Scots pine. Commun. Inst. For. Fenn. 80(2). pp. 1-105. Venäläinen. M.. Harju, A., Nikkanen, T., Paajanen, L., Veiling, P., and Viitanen. H. 2001. Genetic variation in the decay resistance of Siberian larch (Larix sibirica Ledeb.) wood. Holzforschung, 55(1): 1-6. Viitanen, H.. Paajanen. L., Nikkanen. T., and Veiling, P. 1998. Decay resistance of Siberian larch wood against brown rot fungi. Part 2. The effect of genetic variation. International Research Group on Wood Preservation. Stockholm. Doc. IRG/WP 98-10287. Wilkes. J. 1991. Heartwood development and its relationship to growth in Pinus radiata. Wood Sci. Technol. 25: 85-90. Zabel. R.A., and Morrell. J.J. 1992. Wood microbiology: decay and its prevention. Academic Press Inc., San Diego, Calif. Zobel. 8.J., and van Buijtenen, J.P. 1989. Wood variation, its causes and control. Springer-Verlag. Berlin. Heidelberg. Scand. J. For. Res. 17: 199-205, 2002 ® Taylor & Francis Taylor & Francis Croup © 2002 Taylor & Francis. ISSN 0282-7581 Genetic Parameters Regarding the Resistance of Pinus sylvestris Heartwood to Decay Caused by Coniophora puteana ANNI M. HARJU and MARTTI VENÄLÄINEN Finnish Forest Research Institute, Punkaharju Research Station, FI-58450 Punkaharju , Finland INTRODUCTION The best way to protect wooden structures against fungal attack is to use construction techniques which ensure that the wood remains dry. If there is a risk of wood being exposed to moisture for longer periods, it would be both an environmentally and ecologically sound option to use wood that is naturally durable against decay. In Finland, the heartwood of old growth Pinus sylvestris L. and an extract made from this material (pine tar) have been traditionally used in structures exposed to the risk of decay (e.g. Löyt tyniemi 1986). To increase the production and quality of P. sylves tris heartwood, it is essential to know the variation in the rate of formation and the properties of heartwood between different stands and between the trees within the stands. If the variation is to a great extent of genetic origin, it could be used by means of tree breeding, e.g. recurrent selection. Several studies on P. sylvestris suggest that it would be possible to improve the rate of heartwood formation and its chemical properties by breeding (Fries & Ericsson 1998, Ericsson & Fries 1999, Fries et al. 2000, Ericsson et al. 2001). Studies by Schmidtling & Am burgey (1982) on Pinus taeda L. and Venäläinen et al. (2001) on Larix sibirica Ledeb. also showed moderate Harju, A. M. and Venäläinen, M. (Finnish Forest Research Institute, Punkaharju Research Station, FI-58450 Punkaharju, Finland). Genetic parameters regarding the resistance of Pinus sylvestris heartwood to decay caused by Coniophora puteana. Received June 8, 2001. Accepted October 15, 2001. Scand. J. For. Res. 17: 199-205, 2002. Genetic variation in the durability of Pinus sylvestris L. heartwood to a brown rot fungus, Coniophora puteana, was studied using an in vitro decay test. Juvenile heartwood was sampled from 33-yr-old half-sib families growing in a progeny test and from their mothers in a clonal archive. The narrow-sense heritability for the heartwood weight loss was 0.37, and the coefficient of additive genetic variation was 28%. Heritability estimated by the regression of the offspring on mothers was 0.29, and the coefficient of genetic prediction was 0.24. These results indicated that it would be possible to improve the decay resistance of P. sylvestris heartwood by direct selection. According to the genetic correlation (rA = 0.36), selecting for heartwood density would result in an unfavourable response in weight loss caused by C. puteana. However, it appears that unknown environmental factors, which increase heartwood density, also decrease the heartwood weight loss (rE = 0.56). This result emphasizes the need for better understanding of the relationships among wood density, decay fungi, and environ mental factors. Key words: basic density, brown rot, clonal archive, genetic correlation, heritabil ity, offspring-parent regression, progeny test, Scots pine, wooden structures. Correspondence to: A. M. Harju, e-mail: anni.harju@metla.fi broad-sense heritabilities (0.22 and 0.39, respectively) in durability against brown-rot decay fungus in a laboratory test. This group previously carried out a study on P. sylvestris, and found very low narrow-sense heritabil ities for both sapwood and heartwood durability (0.04 and 0.07, respectively), although the coefficient of additive genetic variation for the heartwood durability was relatively high (19%) (Harju et ai. 2001). In that progeny test several of the studied trees were slightly curved. The curved growth habit was induced by an unspecified and experimentally uncontrolled factor which occurred irregularly in the experiment. It was speculated that the reaction wood present in the curved stems degraded at a different rate to the "normal" wood. As a result of the uncontrolled and irregular occurrence of the reaction wood, the residual variance for heartwood durability became very high compared with the additive genetic variance, resulting in low narrow-sense heritability. The aim of the present study was to evaluate further the possibility of improving the heartwood decay resistance of P. sylvestris by tree breeding. To eliminate the disturbing effects of the curved growth habit of the stems, the wood material for the present study was obtained from a P. sylvestris progeny trial with straight stems. The necessary genetic parameters, narrow-sense 200 A. M. Harju and M. Venäläinen Scand. J. For. Res. 17 (2002) heritability (h 2 ) and the coefficient of genetic varia tion (CVa), were estimated for heartwood weight loss, which is inversely related to decay resistance. Heartwood weight loss data were obtained from a laboratory test in which heartwood samples were exposed to a primarily cellulose-degrading brown rot fungus, Coniophora puteana (Schum. ex Fr) Karst., which causes decay in wooden constructions exposed to moisture and water (Viitanen & Ritschkoff 1991). Heritability was estimated as parent-offspring re gression and by the coefficient of genetic prediction. The genetic parameters for heartwood weight loss were compared with the parameters for the wood density, which were well known, and to the parame ters for the amount of heartwood. MATERIALS AND METHODS Study populations, sampling and preliminary analysis of increment cores The primary set of P. sylvestris heartwood samples, consisting of 5-mm diameter increment cores, was obtained from a progeny trial comprising half-sib families originating from open pollination of selected plus trees. The trial (no. 310/1 according to the Finnish Forest Genetic Register) is located in Korpi lahti in central Finland (62°11' N, 25°23' E, on 135 m elevation) and is owned by the Finnish Forest and Park Service. The trial was planted in 1968 in eight randomized complete blocks with square plots of 16 2-yr-old, bare-rooted seedlings at a spacing of 2 x 2 m. The trial was thinned in winter 1989-1990 to a density of eight trees per plot. Twenty-six families of similar east Finnish origin were used to estimate the narrow-sense heritabilities. Core samples were taken from two dominant straight-stemmed trees per plot from each block, in total 16 trees per family. The increment core samples were drilled during a 1 week period at the end of August 1999. One incre ment core was sampled from a random direction at the midpoint between two whorls on each tree. As the diameter of the heartwood was required to be at least 40 mm for the decay test, the sampling height varied to meet this criterion. The average sampling height was 91 cm and it ranged from 60 to 128 cm above ground level. As the samples were not taken from a specific height (e.g. breast height) this resulted in an approximately equal number of annual rings in the increment cores sampled from individual trees. The mean number of annual rings was 28 (SD = 0.9). The cores were drilled through the pith for a distance of a few centimetres on the opposite side of the pith. The transition point between the sapwood and heartwood was marked according to visually observed moisture differences immediately after the increment core was sampled. This made it possible to count the number of sapwood and heartwood rings later in the labora tory. The samples were stored in sealed plastic tubes at a temperature of 5°C. The total number of increment cores was 413 (samples from three trees were missing). A second set of increment cores, which was used to estimate offspring-parent regression, was obtained from a clonal archive located in Punkaharju in south eastern Finland (61°48' N, 29°20' E, on 85 m eleva tion, established in 1968-1971). Increment cores were taken from two grafts of each of 20 clones during a 3 day period at the beginning of September 1999. The clones were grafted from the mother trees of the 20 half-sib families in the progeny test described above. Because the grafts were slightly curved, two incre ment cores were bored at right angles to each other and at different heights. This was done to make the clone-wise measurements more precise. The lower sampling height was 98 ± 11.4 cm (mean + SD) and the upper one was 133 + 8.2 cm. Cores were sampled and stored in the same way as those used for the progeny test. Laboratory experiment A 40 mm section was cut from the heartwood part of each increment core. However, in a few cases only a 30 mm section was cut, because the aim was to obtain pure heartwood. The sections were dried at 60° C for 48 h, after which they were cooled in a desiccator and weighed to an accuracy of 1 mg. The basic density of the wood specimen (mg cm -3) was calculated from the dry weight of the sample and its fresh volume (a cylinder with a diameter of 5 mm and length of 40 or 30 mm). The weighed samples were packed into paper bags and sterilized at VTT Chemical Technol ogy using a radiation dose of between 25 and 50 kGy from a radioisotope '"Co source. The decay resistance was studied at VTT Building Technology using a malt agar plate decay test, which is a modification of the standardized EN 113 (Eu ropean Standard 1996) as described by Viitanen et ai. (1998) and Venäläinen et ai. (2001). The cylindrical increment core sections, from which the wood extrac tives had not been removed, were placed on a pure culture of a brown rot fungus, C. puteana (Schum. ex Resistance of P. sylvestris to decay 201 Scand. J. For. Res. 17 (2002) Fr.) Karst. (strain BAM Ebw. 15), growing on agar in Petri dishes. The samples were placed on a glass rack so that they were not in contact with the agar. A random set of four wood specimens was placed in each Petri dish. The incubation time was 8 weeks, after which the samples were dried at 60° C for 48 h and reweighed. The weight loss during the experiment was expressed in absolute terms as mg cm -3 , and in relative terms [(weight loss/start weight) x 100%]. Statistical analysis On the individual tree level, a measured phenotypic value (P) of a trait is assumed to be the sum of the additive genetic effect (A) and the independent envi ronmental effect ( E). E also includes the remaining genetic effects that are independent of A. Thus, P A + E. Furthermore, the phenotypic variance at the population level is assumed to be composed of ge netic and environmental components, a 2 P =a 2 + o\. The narrow-sense heritability is estimated using h 2 = ff 2 A l,