THE BEHAVIOUR AND IMPACTS OF ARMILLARIA OSTOYAE INMATURE STANDS AND PLANTATIONS IN THE SHUSWAP REGION OFBRITISH COLUMBIAbyALEX J. WOODSB.Sc., University of Alberta, Edmonton, Alberta, 1989A THESIS SUBMI’FIED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Forest Sciences)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMay 1994Alex 3. Woods, 1994in presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission._________________________Department ofThe University of British ColumbiaVancouver, CanadaDate Jt1N. -‘4DE-6 (2/88)1).ABSTRACTArmillaria ostoyae causes considerable loss in forest productivity in bothimmature and mature stands within the Interior Cedar Hemlock (ICR) and the InteriorDouglas-fir (IDF) zones of the southern interior of British Columbia. Two studiesconcerning the impacts of this pathogen where conducted near Salmon Arm, B.C.; onewas within four plantations age ten to twenty-five years on Larch Hills, and the otherwas within mature stands on Hunter’s Range and Larch Hills.In the plantation study the relationship between the levels of A. ostoyaeinfection in mature tree stumps and the regeneration was examined. The evidence ofpast A. ostoyae infection in stumps remains visible on the inner bark for at least thirtyyears. This evidence may be used to estimate the extent of the disease in the formerstand. The relationship between A. ostoyae in stumps and A. ostoyae-caused mortalityin regeneration was significant, though not strongly. Three measures of stumpinoculum were compared: the proportion of stumps infected, the number of stumpsinfected, and the basal area of stumps infected. The number of stumps infected wasmost closely associated with the proportion of regeneration infected. The relative ratesof infection incidence were compared among the eight regeneration species present inthe four plantations. A quantitative means of comparing the incidence of infectionamong species was developed. The number of infected Douglas-fir trees was chosen asa standard measure of disease incidence for each plantation. The incidence of infectionin the other seven species were then compared to the Douglas-fir standard. Theprobability of a young tree becoming infected with A. ostoyae did not increase as thedistance from an infected stump was reduced. Brushing in one of the twenty-five yearold plantations significantly increased the mortality caused by A. ostoyae.The second study was concerned with the impacts of A. ostoyae in maturestands approximately 120 years old. The relative rates of incidence of A. ostoyaeiiiinfection were compared between species. The incidence of infection for Western larch(Larix occidentalis) was no less than that for Douglas-fir (Pseudotsuga menziesii). Theranking of tree species susceptibility may depend more on site than on inherentdifferences among species. An A. ostoyae severity rating system was developed. Thissystem assigned plots a rating based on the proportion of conifer trees infected out ofthe total number of conifers in the plot. This severity rating was then used in analysesto test the relationship between A. ostoyae severity and a variety of site characteristics,including elevation, logging disturbance, site index, and biogeoclimatic siteclassification. Of these characteristics, past logging disturbance was mostly closelyassociated with high levels A. ostoyae severity. The relationship between thebiogeoclimatic ecological classification system and A. ostoyae incidence and severitywas examined. The ICR zone had significantly more A. ostoyae infection than the IDFzone. More detailed analyses using site units within both the ICH and IDF zones didnot indicate any significant relationships. The A. ostoyae severity rating was alsocompared to timber volumes. There was a significant relationship between A. ostoyaeseverity and conifer volume in 120 year old stands in the ICH zone. The most severelyinfected plots had significantly less conifer volume than the less heavily infected areas.There was no significant relationship between A. ostoyae severity and conifer volumein the IDF zone. However, there was a clear trend towards lower conifer volumes withincreasing A. ostoyae severity.ivTABLE OF CONTENTSAbstractTable of Contents ivList of Tables viiList of Figures xAcknowledgments xiChapter 1 General Introduction 11.1 The Genus Armillaria 31.1.1 Parasitic and Saprophytic Phases 31.2 The Disease 51.2.1 Inoculum Potential 51.2.2 Inoculum Potential and Disease Dynamics 51.2.3 Infection 71.2.4 Hosts 91.2.5 Disease Recognition, Symptoms and Signs 101.3 Disease Detection and Assessment 131.3.1 Detection 131.3.2 Assessment 131.4 Summary 15Chapter 2 Arinillaria ostoyae Development in Plantations, Following theHarvest of Infected Stands 172.1 Introduction 172.2 Methods 212.2.1 Study Areas 212.2.2 Sampling Design 222.2.3 Sampling Procedure 222.2.4 Data Analysis 25V2.3 Results 362.3.1 Stump Results 362.3.2 Regeneration 462.3.3 The Relationship Between Past Levels of Armillariaostoyae in Mature Stands and Present Levels of theDisease in Plantations Established on the Same Sites 562.4 Discussion 612.4.1 Comparison of the Relative Rates of Armillariaostoyae Infection Incidence Among Conifer Speciesto a Douglas-fir Standard 612.4.2 The Impacts of Armillaria ostoyae on SpeciesComposition and Succession Within the ICH Zone 632.4.3 The Relationship Between Past and Present Levelsof Armillaria ostoyae on a Site 662.4.4 The Proximity of Infected Regeneration to InfectedStumps 692.4.5 The Impacts of Bnishing Stands Infected WithArmillaria ostoyae 71Chapter 3 Armillaria ostoyae Distribution and Severity in the ICH and IDFBiogeoclimatic Zones 753.1 Introduction 753.2 Methods 783.2.1 Study Area 783.2.2 Sampling Design 783.2.3 Sampling Procedure 793.2.4 Data Analysis 833.3 Results 873.3.1 Individual Tree Results 873.3.2 Plot Results 953.4 Discussion 1033.4.1 The Ranking of Conifer Species Based onIncidence of Armillaria ostoyae Infectionsin the ICH and IDF Zones 103vi3.4.2 The Influence of Prior Disturbance on theExpression of Armillaria ostoyae in the ICH andIDF Zones 1063.4.3 The Influence of Elevation on Root DiseaseExpression 1073.4.4 The Relationship Between Armillaria ostoyaeDistribution and the Biogeoclimatic SiteClassification in the Southern Interior of B.C. 1073.4.5 The Relationship Between Armillaria ostoyaeSeverity and Stand Productivity in Terms ofTimber Volume 1093.4.6 The Relationship Between Phellinus weiriiand Armillaria ostoyae 1103.4.7 The Value of Surveying for Armillaria ostoyae inMature Stands in the ICH and IDF Zones 111Chapter 4 Conclusions 112Bibliography 115Appendices 120viiLIST OF TABLES1 Ranking of tree species susceptibility to Armillaria ostoyae in thesouthern interior of British Columbia (from Morrison et al. 1991a) 92 Forest management history records for the four plantations 223 Relative rates of infection compared between lodgepole pine and Douglas-fir regeneration 314 Observed and expected numbers of lodgepole pine and Douglas-firwithin the four plantations 325 Percentage of total stumps by infection status in each species for allfour plantations 366 Percentage of total stumps positively identified by infection status ineach species for all four plantations 377 Comparison of the proportion of stumps infected pre- and post-harvest,clean and no-evidence among the three major species in all fourplantations combined (proportions expressed in percent of total inthat species) 378 Comparison of the percent of stumps infected with Annillaria ostoyaein all four plantations to that of the surrounding mature trees 389 Percentage of total stumps in each of the three major species in eachorigin class 4010 Percentage of total stumps of the three major species in each of ten10cm diameter classes all plantations combined 4111 Percentage of stumps infected with Armillaria ostoyae out of the totalidentified stumps (excluding no evidence stumps) for each plantationin percent 4312 Species composition of the stumps in each of the four plantations inpercent 4413 Mean stumps basal area for each plantation by disease status (m2/ha) 4514 Number of conifer trees by species tallied/ha in each of the fourplantations 4815 Ranking of conifer species compared to a Douglas-fir standard on thebasis of the incidence of Arinillaria ostoyae infection in plantationsage 10-25 years 4916 Chi-square values from the comparison of the proportions of treesinfected with Annillaria ostoyae among eight conifer species in fourplantations aged 10 to 25 years 50viii17 Conifer regeneration species distribution in percent by diameter class(all plantations) 5118 Armillaria ostoyae severity distribution by diameter class and plantation(% of total in D-class) 5119 Percentage of stems infected with Armillaria ostoyae out of the total numberof stems in each diameter class for the four major regeneration species 5220 Percentage of total stems in each diameter class by plantation 5421 Percentage of total conifer trees in each plantation killed by or dyingfrom Armillaria ostoyae or killed by other causes 5422 Distribution of plots by increasing frequency of infection for all fourplantations 5623 Correlation coefficients resulting from the correlation of stump andregeneration variables for each plantation individually and for all fourplantations combined compared to critical values (a = 0.05 and 0.01) 5724 Effects of plot size on relationships between the number of infectedstumps and the number of infected regeneration 6025 A comparison of the two 1969 plantations 7226 Percentage of total trees infected with Armillaria ostoyae in percentby species by zone (Hunter’s Range) 8827 Percentage of total trees infected with Armillaria ostoyae comparedamong five major species along with the percent of trees infected withother diseases in percent by species by zone (Larch Hills) 8928 Comparison of the percentage of total trees infected with Armillariaostoyae in each of the major species between the Larch Hills and Hunter’sRange study areas 9029 Comparison of the proportions of the five major species in percentbetween the Larch Hills and Hunter’s Range 9130 Ranking of conifer species compared to a Douglas-fir standard on thebasis of the incidence of Armillaria ostoyae infection in mature stands(approx. 120 years) within the ICH and IDF zones of Hunter’s Rangeand Larch Hills 9131A Percentage of trees infected in percent for each species in each site unitin the ICH zone 923 lB Percentage of trees infected in percent for each species in each site unitin the IDF zone 9332 Comparison of the 1992 and 1993 methods of data collection regardingthe number of infections found, and the percent of total trees missedusing the 1992 method 94ix33 Plot distribution by site unit, disease status, and disturbance history 9534 ANOVA results for Armillaria ostoyae severity vs disturbance class in theICR and IDF zones 9735 ANOVA results for Armillaria ostoyae severity vs biogeoclimatic site unitin the ICH and IDF zones 9936 ANOVA results for Armillaria ostoyae severity vs conifer volume in theICH and IDF zones individually and combined 10037 The relationship between Armillaria ostoyae severity and conifer volumewith the number of pio in each severity class and the associated meanconifer volume/plot (m /ha) for the ICH and IDF zones 10138 ANOVA results for Armillaria ostoyae severity vs paper birch volume inthe ICR and IDF zones 10139 The relationship between Armillaria ostoyae severity and birch volumewith the number of plfts in each severity class and the associated meanbirch volume/plot (m /ha) for the ICH and IDF zones 10240 Test of independence between Armillaria ostoyae and Phellinus weiril inthe ICH and IDF zones 103xLIST OF FIGURES1 Stump distribution among plantations by disease status 422 Distribution of conifer regeneration by species in all four plantations 473 The relationship between the number of stumps infected pre-harvest andthe proportion of conifer regeneration infected in the 1969b plantation 584 Arinillaria ostoyae severity compared between disturbed and undisturbedplots (ICH and IDF combined) 965 The relationship between Armillaria ostoyae severity and plot elevationin the ICH and IDF zones 986 The relationship between Armillaria ostoyae severity and biogeoclimaticsite unit in the ICH and IDF zones 99ACKNOWLEDGMENTSI would like to thank my supervisory committee, in particular Dr. B. J. van derKamp, Dr. P. Marshall, Dr. D. Morrison, and H. Merler, for their support and input.Funding for this research was provided by the Canada - British Columbia PartnershipAgreement on Forest Resource Development, FRDA II. I would like to thank myparents for providing me with room and board over the two field seasons. I would alsolike to thank Bess for her help and patience with data collection, and Rory for his helpwith computer software. Most of all I would like to thank my wife, Jane, who hashelped so much with all aspects of this research with her assistance, patience andsupport.xi1CHAPTER 1 GENERAL INTRODUCTIONArmillaria ostoyae (Romagn.) Herink, which causes Armillaria root disease, is one ofthe most serious forest pathogens in southern British Columbia and the Pacific Northwest states(Morrison et a!. 1985, Wargo and Shaw 1985). Annillaria ostoyae reduces timberproductivity by reducing tree growth, predisposing trees to attack by insects and otherpathogens, and by causing direct mortality.Armillaria ostoyae has a parasitic phase where the pathogen can aggressively colonizenew hosts, and a saprophytic phase where the fungus utilizes the colonized food base forgrowth (Morrison 1981). In the saprophytic phase, A. ostoyae can survive for decades in thelarge woody portions of its host following host mortality. The persistence of the fungus onsite results in long term losses in timber production on sites infested by this pathogen.Armillaria ostoyae behaves quite differently on the coastal portions of its rangecompared to the interior. West of the coast mountain range in British Columbia A. ostoyae isprimarily a disease of young plantations (Morrison 1981). In this area, mortality caused bythe disease is normally restricted to trees less than 25 years of age, although trees under severestress may be killed at any age. Annillaria ostoyae impacts are more severe east of the coastmountain range in the southern interior of British Columbia as the pathogen continues to killtrees throughout their rotation (Morrison 1981). In the southern interior, the impacts of A.ostoyae are most pronounced in the Interior Cedar Hemlock (ICH) biogeoclimatic zone, butare also felt in the Interior Douglas-Fir (IDF), Montane Spruce (MS), and Sub-Boreal Spruce(SBS) zones (Morrison et a!. 1991a). The ICH zone has the greatest diversity of tree speciesof any biogeoclimatic zone in British Columbia. Armillaria ostoyae is an integral componentof this zone and is at least in part responsible for the species diversity found in theseecosystems. Morrison (1981) suggests A. ostoyae accelerates succession in this zone by killingpioneer species such as Douglas-fir (Pseudotsuga menziesii Mirb. Franco) and lodgepole pine(Pinus contorta Dougl. ex Loud.), allowing the shade tolerant western hemlock (Tsugaheterophyla (Rafn.) Sarg.) and western redcedar (Thuja plicata Donn ex D. Don) to move in.2These species are not distinctly less susceptible to A. ostoyae, but appear to be more tolerant(Morrison 1981). Armillaria ostoyae may also delay succession in this zone by maintainingheavily infested sites in early seral deciduous species. Deciduous species such as paper birch(Betulapapyrifera Marsh.), when young, are the most tolerant species to this disease.Annual losses of timber due to A. ostoyae within the ICH zone are estimated to be 105000 cubic meters (Taylor 1986). The ICH zone is the most productive forest zone in theinterior of British Columbia and is second only to the Coastal Western Hemlock (CWH) zonefor productivity in all of Canada (Ketcheson et a!. 1991). Potential gains in productivity fromreducing A. ostoyae impacts are therefore high in the ICH zone. This research projectconcentrated on the behavior of A. ostoyae within the ICR zone.The first objective of this study was to develop a better understanding of thedevelopment of A. ostoyae in plantations established on sites that were previously infested withthe disease. A second objective was to determine which factors out of biogeoclimatic sitetype, site index, disturbance history and elevation have the greatest influence on thedistribution and intensity of A. ostoyae within the southern interior of British Columbia. Theimpact of A. ostoyae severity on timber volume was also examined.This thesis consists of four chapters. Chapter 1 is a general introductory chapter whichincludes a brief literature review of the background biology of A. ostoyae as well as a reviewof literature relevant to Chapters 2 and 3. Chapter 2 covers an examination of the relationshipbetween past levels of A. ostoyae infection in mature stands and the subsequent levels ofinfection in plantations once those mature stands are harvested and regenerated. Chapter 3covers an examination of the relationship between A. ostoyae infection severity and several sitefactors. These site factors included biogeoclimatic site classification, elevation, prior loggingdisturbance, and site index. The impact of various levels of A. ostoyae infestations on timbervolume was also examined. Chapters 2 and 3 both contain four sections: introduction,methods, results and discussion. Chapter 4 consists of the conclusions drawn from theexaminations conducted in Chapters 2 and 3.31.1 The Genus ArmillariaArmillaria ostoyae is one of approximately 36 species that belongs to the genusArmillaria within the order Agaricales (Watling et at. 1991). Several features of theArmillaria genus are distinctive from other members of the order. These features include theapparent diploid nuclei in vegetative mycelium, the presence of both parasitic and saprophyticcapabilities, and the production of rhizomorphs. Fruiting bodies (basidiomes) are essential forthe complete description and naming of species (Watling et at. 1991). Two species ofArmillaria occur throughout the southern interior of British Columbia, A. sinapina and A.ostoyae (Morrison et at. 1985). Species of Armillaria occur worldwide. For example A.borealis occurs in northern Europe and Russia, A. luteobubalina in Australia, and A.montagnei in South America (Watling et at. 1991). Armillaria ostoyae and A. gallica are twoof the most widely distributed species and are circumpolar within the northern hemisphere(Guillaumin et at. 1989).1.1.1 Parasitic and Saprophytic PhasesArmillaria spp. have both saprophytic and parasitic phases in their life cycle. Withinthe genus, species differ widely in pathogenicity. Highly pathogenic species such as A.ostoyae survive saprophytically in the hosts they have killed during their parasitic phase.Armillaria ostoyae, however, appears to be a weak saprophytic competitor. Unless the fungushas colonized host tissue prior to host death, A. ostoyae is considered incapable of replacingother fungi in dead stump tissue (Garrett 1970). Thus, stumps that exhibit signs of A. ostoyaeinfection were infected prior to the death of the stump. Stump tissues, although greatlyweakened, remain alive for a year or more following cutting which allows A. ostoyae torapidly colonize the entire stump in the absence of competition from obligate saprophytes. Cutstumps, therefore, provide A. ostoyae with large sources of inoculum and give the fungus anadvantage over competing fungi. Trees killed by A. ostoyae also represent large sources ofinoculum, however, the fungus has to exert much more energy in order to create that food4supply. In this case, A. ostoyae not only has to infect the host, but also has to battle with thehost’s natural resistance mechanisms in order to provide the food supply for the saprophyticstage. Whether the increase in A. ostoyae inoculum is the result of logging or from disease-caused mortality, once the inoculum potential is high on a site, the disease may spread rapidlyfrom tree to tree across host root contacts. In mature stands the disease may continue tospread at a high rate until the disease contacts less susceptible hosts or reaches a naturalbarrier.Infected stumps are the initial sources of inoculum in plantations. Very young treesgrowing in contact with infected stumps are quickly killed. These small trees do not contributeto the spread of the disease because of lack of root contacts and their low inoculum potential.The inoculum potential of the colonized stumps decreases with time as A. ostoyae and othersaprophytic organisms consume the stump tissue. Consequently, in order for the disease tocontinue to spread in plantations, the young trees must reach a size where root contactsbetween trees are common and where newly killed trees represent significant amounts of foodbase for the pathogen.Weak pathogens such as A. sinapina probably exist for the most part as saprophytes(Korhonen 1978, Rishbeth 1985, Wargo and Shaw 1985). Field observations of A. sinapinasuggests that this species is not capable of attacking healthy conifers (Morrison et al. 1991a).Weak pathogenic species do have the ability to act as a facultative parasites on stressed orunhealthy hosts (Kile 1980, Rishbeth 1985, Wargo and Shaw 1985), as do highly pathogenicspecies such as A. ostoyae. The ability of the Armillaria genus in general to take on bothparasitic and saprophytic roles has undoubtedly contributed to the success of the genus aroundthe world.51.2 The Disease1.2.1 Inoculum PotentialAll Armillaria species survive saprophytically in woody substrates in soil and mayexhibit varying degrees of pathogenicity (Redfern and Filip 1991). The survival and vigor ofan Armillaria clone depends primarily on its inoculum potential. Garrett defined (1960) andthen redefined (1970) inoculum potential as “the energy of growth of a parasite available forinfection of a host at the surface of the host organ to be infected”. Inoculum potential isdependent upon the surface area of fungus in contact with the host, the vigor of the invadinghyphae, and environmental effects on the fungus (Redfem and Filip 1991).Inoculum potential of Armillaria spp. is also influenced by the distance between thehost and an inoculum source. Inoculum potential is maximized where healthy roots andinoculum are in contact. Where gaps are bridged by rhizomorphs, the inoculum potentialdiminishes with increasing distance between inoculum source and host (Redfern and Filip1991).1.2.2 Inoculum Potential and Disease DynamicsInoculum potential of A. ostoyae on a site may be greatly increased by forestryoperations such as clear-cutting, selective cutting, and thinning. Harvesting, particularlyselective logging, has led to inoculum build up on many species of conifer stumps in westernNorth American forests (Byler et al. 1990). For the pathogenic species of Armillaria, moreinoculum typically results in more disease (Redfern and Filip 1991).The inoculum distribution and potential on sites dictates to a large extent the dynamicsof the disease. In natural undisturbed forests where A. ostoyae is present, diseased trees occurin groups or as scattered individuals around sources of inoculum (Morrison et al. 1991a). Thedisease may be actively attacking new hosts or be quiescent in lesions on infected tree roots.In these undisturbed forests, the mechanism responsible for A. ostoyae switching from a6quiescent state to an active state is not well understood. Stresses caused by drought, insectattack or other diseases, such as white pine blister rust may play a role. In managed forests,the transition to the active, aggressive disease state is obvious. Following cutting, the stump israpidly colonized from active infections or quiescent lesions. Armillaria ostoyae then producesa flush of rhizomorphs (Morrison et al. 1991a). The inoculum load is then at its maximumbecause the food base is freshly colonized (Morrison et a!. 1991a).In clearcut areas, there is usually a delay of four to five years following planting beforesignificant mortality is observed. This is the length of time required for the fungus to fullycolonize the stumps and for A. ostoyae rhizomorphs to contact the young tree roots. Mortalityinitiated by root contacts with an inoculum source does not appear until age ten to fifteen(Morrison et a!. 1991a). Once there is root closure in a stand, A. ostoyae has the potential tospread until a physical or non-host barrier is encountered (Morrison et a!. 1991a). Hagle andGoheen (1988) refer to a wave pattern of mortality that takes place as stands on infested sitesreach pole size. Most of the young trees are killed before they reach a merchantable size andare replaced by abundant regeneration. Mortality then slows until some trees reach pole sizeand their root systems become large enough to provide the inoculum to fuel another wave ofmortality. The primary difference between the wave pattern in undisturbed root diseaseinfected stands and infected clearcut stands is that the cutting sets the time of the wave bycreating an immediate food base and stimulating the regeneration that will feed the next wave(Hagle and Goheen 1988).The conflict between stand management goals and A. ostoyae is most clearlydemonstrated in partially cut stands. If the pathogen is present in well stocked stands, theimpacts can be high. Well stocked and overstocked stands are the most logical candidates forpartial harvests. Trees in these stands have a high probability of root contact. As previouslymentioned, once an infected tree is harvested its entire root system is colonized. In partial cutoperations, “leave” trees are then exposed to large sources of inoculum. Repeated partialharvests every fifteen to twenty years often result in severely infected sites on which few treesreach merchantable size (Morrison et a!. 1991a). The negative impacts of partial cutting on7timber supply in A. ostoyae infected stands is even greater when the more A. ostoyae toleranttree species are selectively harvested (Hadfield 1984, Hagle and Goheen 1988). In the PacificNorthwest states, the combined effects of succession, fire control and selective harvesting hasresulted in the removal of western white pine (Pinus monticola Dougi. ex D. Don), ponderosapine (Pinus ponderosa Dougi. ex Laws.) and western larch (Larix occidentalis Nutt.) leaving apredominance of more susceptible species such as Douglas-fir and grand fir (Abies grandis(Dougi. ex D. Don) Lindi). Similar practices in the southern interior of British Columbia,involving the selective removal of western white pine, and larger Douglas-fir individuals haveprobably had similar effects on overall timber supply.Fire also has an influence on the levels of A. ostoyae infestation. Fire may directlyaffect the activity of the pathogen in forests through reduction of inoculum or indirectlythrough stress effects on the fungal mycelium which leads to natural biocontrol (Reaves et al.1990). On the other hand, fire may lead to a flush of A. ostoyae activity similar to that foundin cut over areas. By killing infected trees, fire may allow the fungus to rapidly colonize theremaining portions of the tree leading to a build-up of inoculum on the site. Fire alsoinfluences the species which regenerate sites, favoring the early seral species such as Douglas-fir and lodgepole pine. These early seral species tend to be more susceptible to A. ostoyae.The success of A. ostoyae infections depend to a large extent on the inoculum potential. Inorder to control the disease, therefore, the inoculum potential on a site must be reduced.1.2.3 InfectionInitial infection of a host occurs either through rhizomorph contact or through host rootcontact with an inoculum source. Recent work in Europe and North America suggests thatrhizomorph production is more important in the less pathogenic species of Armillaria. Themost pathogenic Armillaria species tend to spread primarily through root contacts (Gregory1985, Morrison 1989).8Penetration of host root bark by rhizomorphs is both a chemical and a mechanicalprocess. The processes vary depending on the host’s root surface morphology (Thomas1934). Root surface morphology typically changes with age, being smooth when the root isyoung and growing progressively rougher over time. In young roots a rhizomorph becomesattached initially by the hardening of the mucilaginous substance which covers its growing tip.Then, single hyphae developing from the rhizomorph tip penetrate the outer layer of cork cellsand anchor the rhizomorph to the root (Morrison et at. 1991b). The rhizomorph thenpenetrates through the cork layer by exerting pressure while at the same time producingsecretions which disrupt the cork cells.In older, scaly roots, rhizomorphs may penetrate the bark scales and develop infectionwedges beneath them. In the area surrounding the infection wedge, cell walls turn brown andcell contents are disrupted (Woeste 1956).Infection may also occur at root contacts between healthy and infected trees. Zeller(1926), based on work on apple trees, suggested that infection of the new host begins whenhealthy bark contacts toxic substances produced by A. mellea in the diseased root. It has beensuggested (Morrison et a!. 1991b) that conifers may become infected in a similar manner.Mycelial fans of A. ostoyae first grow in a root’s outer bark. As the area of infected barkincreases, the mycelial fans penetrate to the inner phloem and cambium.Mycelial fans act as a unit and continue to kill host tissue in the cambial zone,eventually girdling the root. Once the root is girdled, A. ostoyae rapidly colonizes thecambium of the remaining distal portions of the root. The fungus also spreads proximallyfrom the point of infection towards the root collar and onto other primary roots. The locationof infections is an important factor in disease development. The closer the initial infection isto the root collar, the more likely that the tree will be girdled and killed (Morrison et at.1991a). Shaw (1980) found that the most lethal attacks on sapling and pole-sized ponderosapines occurred high on the tap roots or on the root collar. Infections on lateral roots rarelyresulted in lethal infections for this species. After a host is killed, its entire root system israpidly colonized. Thus, the spread of the pathogen through root contacts can occur quite9quickly, particularly in fully stocked stands. Shaw and Roth (1976) estimated the rate ofspread of A. nwllea in a ponderosa pine stand to be 1 rn/year. Van der Kamp (1992) estimatedthe rate of A. ostoyae spread in a 110 year old interior Douglas-fir stand to be 0.22 mlyear. Acolonized root system may contain viable Annhllaria spp. for up to 40 to 70 years afterharvest, depending on the size of the stump and other environmental factors (Kile 1980b,Rishbeth 1972b, Shaw 1975).1.2.4 HostsArmillaria species in general have a very broad host range. The factors whichdetermine host preferences in natural stands are not well understood. In the southern interiorof British Columbia, A. ostoyae attacks all tree species and some shrubs, although conifers areits principle host (Morrison et a!. 1991a). The forests of the southern interior of B.C. are veryspecies diverse. Conifer species found in this area include Douglas-fir, lodgepole pine,western larch, western redcedar, Engelmann spruce (Picea engelmanni Parry ex Engel.),western hemlock, western white pine, ponderosa pine, subalpine fir (Abies lasiocarpa (Hook.)Nutt.), paper birch, and trembling aspen (Populus tremuloides Michx.) among others.Morrison et a!. (1991a) suggested the following ranking of susceptibility among tree species inthe southern interior of British Columbia (Table 1).TABLE 1. Ranking of tree species susceptibility to Armillaria ostoyae in the southerninterior of British Columbia (from Morrison et a!. 1991a)Susceptible: Abies sp.Douglas-firspruceModerately Susceptible: lodgepole pinehemlockcedarTolerant: ponderosa pine?birchResistant: larch10No tree species have shown resistance to A. ostoyae before the age of fifteen, with thepossible exception of paper birch (Merler, B.C. MOF, unpublished data). Ar,nillaria ostoyaecan have a significant influence on stand composition due to the varying degrees ofsusceptibility in host species. Within the ICH zone, western redcedar, and paper birch oftenreplace Douglas-fir as the principle species in stands infected with A. ostoyae. Although bothconifer species exhibit similar susceptibility to initial infections of A. ostoyae, westernredcedar tends to cope with infection better than Douglas-fir. Western redcedar is often ableto surround infections with callus tissue preventing A. ostoyae from girdling and killing thetree. Douglas-fir is less successful at combating this pathogen and, once infected, is morerapidly girdled and killed. Western redcedar acts as an inoculum reservoir for A. ostoyae on asite, and prevents the disease from disappearing due to lack of inoculum. In the most severelyinfested areas, even western redcedar cannot survive, leaving paper birch and brush species todominate the site.1.2.5 Disease Recognition, Symptoms and SignsTrees infected by A. ostoyae may exhibit some or all of the following symptoms,depending on tree age and location of infection: reduction of shoot growth, changes in foliarcharacteristics, stress induced reproduction, basal resinosis and death. The most widelyaccepted hypothesis describing the physiological basis for A. ostoyae symptom developmentinvolves the physical disruption of the hosts vascular system (Morrison et al. 1991b). Infectedtrees younger than ten years rarely show reduced leader growth prior to death, becausemortality occurs within a year of infection (Hintikka 1974). Older, larger trees may showevidence of prolonged infection through reduced leader growth and rounded crowns.However, van der Kamp’s (1992) work in a 110 year old Douglas-fir stand indicated that theaverage time between the first appearance of above ground evidence of infection (mainly basalresinosis) and tree death was only six years. Foliar symptoms in conifers also differ dependingon duration of the infections. Trees which are killed quickly have red brown foliage which11remains on the tree for approximately a year following death. In trees where the diseaseprogresses more slowly the foliage becomes stunted, chiorotic and sparse throughout the crown(Morrison 1981). In response to advanced infection, usually in the season before death, manyconifers produce a crop of cones that are smaller but may be more numerous than normal(Morrison et at. 1991a). Conifers which usually have resin canals (Pseudotsuga, Picea, LarLrand Pinus), or those which form traumatic resin canals (Tsuga and Abies), may produce resinthat exudes through bark fissures at or above the root collar when attacked by A. ostoyae(Morrison et at. 1991a). Resin exudation does not usually occur above ground until thefungus has reached the root collar. Armillaria ostoyae may be well established beneath thebark of a host before any signs of resinosis is visible on the outside of the bark.Callused lesions are another basal stem indicator which is present in some coniferspecies, including Douglas-fir, western redcedar, western hemlock, western white pine andwestern larch. Infections in hosts of these species may be arrested by a strong host responseonce the trees are more than twenty years old (Morrison et al. 1991a). Armillaria ostoyaelesions can be surrounded by callusing and the spread of infection held in check. Thisphenomenon is most common in western redcedar, where the callusing results in flattenedportions of the bole and eventual trunk fluting. Beneath the bark of the flattened areas is thecharacteristic thangular shaped lesion which is formed above the infected root.Many of these symptoms are not specific to A. ostoyae infections and may be inducedby several other biotic and abiotic factors. Confirmation of A. ostoyae infections requires anexamination of the lower bole, root collar and occasionally the larger lateral roots in closeproximity to the root collar, for specific signs of the fungus.The specific signs used to confirm A. ostoyae infection include mycelial fans,rhizomorphs, and basidiomes. Armitlaria ostoyae may also be further confirmed by culturingthe pathogen from the host (Morrison et at. 1991b).Mycelial fans beneath the bark are the most useful diagnostic characteristic of A.ostoyae in woody species. Mycelial fans may be present early in the infection beneath thebark prior to resinosis and host mortality. Mycelial fans are most abundant in the first few12years following host death. They tend to disappear from the above ground portions of deadhosts over time as a result of competition from other fungi, desiccation, and possiblyconsumption by insects. On conifers, impressions of fans in resin and bark may be present forseveral years after the fans have gone (Morrison et al. 1991a). Under suitable conditions themycelial fan impression in resin and bark of stumps may remain for over thirty years. Thesesigns of past A. ostoyae infections are useful for determining past distribution and incidence ofinfection in cut over areas. Mycelial fans are also useful for determining the cause of callusedbasal lesions.Rhizomorphs of A. ostoyae are initiated on a colonized root system and grow into thesoil when a mycelial fan reaches the bark-soil interface (Morrison 1972). Rhizomorphs arewhite within one centimeter of the growing tip, becoming progressively darker shades ofbrown and eventually black. Rhizomorphs in soil and on the surface of roots are usually 1-3mm in diameter (Morrison 1972). Rhizomorph morphology may be used as a distinguishingfeature to differentiate between species of Armillaria. Two species of Armillaria arewidespread in the southern interior of British Columbia. Armillaria sinapina formsmonopodially branched rhizomorphs which create extensive networks in soil and on rootsurfaces. Armillaria ostoyae forms dichotomously branched rhizomorphs which grow only afew centimeters from the food base (Morrison et at. 1991a).The basidiomes of A. ostoyae are found in clusters on the stem of the host or in the soilabove diseased roots. Basidiomes often occur on or near hosts lacking other above-groundsigns and symptoms (Morrison et at. 1991a). In British Columbia, basidiomes typically formin late August through mid-October when favorable temperatures and precipitation occur.Mature basidiomes of A. ostoyae are 10-20cm tall, with a cream to brown cap 5-10cm widewith a distinct ring on the stem (Morrison et a!. l991a).In conifers, wood with incipient A. ostoyae decay is stained gray to brown, often with awater-soaked appearance (Morrison et a!. 1991a). Later, as A. ostoyae continues todecompose the lignin and cellulose components, the decayed wood becomes yellow brown andstringy, and eventually is reduced to very wet stringy rot, with pale yellow flecks (Williams et13al. 1989). Decayed wood of broadleaved hosts is water soaked and white to yellow, becomingspongy and ultimately distinctly gelatinous (Greig and Strouts 1983).1.3 Disease Detection and Assessment1.3.1 DetectionArmillaria ostoyae may be detected in stands based on its signature (i. e, its diagnosticsymptoms and signs). Stands infested with A. ostoyae may differ considerably in appearance.A large disease center may consist of openings of several hectares containing mostly deciduousand brush species with some young conifer regeneration, surrounded by a perimeter of deadand dying mature conifers. In less severely infected areas, western redcedar may survive asthe sole conifer component making up the majority of the stocking inside the disease center.Throughout the perimeter of a disease center in older stands, stubs, snags and recently killedtrees may be found. Islands of healthy conifers are also often found within disease centerperimeters. These islands or “escapes” occupy areas in which viable A. ostoyae inoculum wasnot present at stand renewal. Such islands usually have symptomatic trees along their edges.Armillaria ostoyae may also be found in very small disease centers of only one or two infectedtrees. These small centers are often diffusely distributed throughout mature stands within theICH zone. The diffuse nature of the infections in some stands makes some forms of surveyingfor A. ostoyae difficult.1.3.2 AssessmentA variety of methods have been devised to detect A. ostoyae and to assess the extentand impacts of the disease. Williams and Leaphart (1978) used large scale colour infraredaerial photographs to estimate the area of root disease centers in northern Idaho. James et al.(1984) used a combination of large scale aerial photographs and existing timber inventory data14to estimate root disease losses in northern Idaho and northwestern Montana. Similar methodsusing large scale aerial photography to assess extent of area infected and resulting timber lossesin the southern interior of British Columbia have not yet been attempted. There are severalpotential problems associated with using these methods in this area. First, the infection centerboundaries themselves are difficult to delineate in many areas due to the diffuse nature ofdisease distribution. Second, even in areas with discrete disease centers, the extent of diseaseoutside those centers is impossible to estimate accurately from photographs. Dubreuil (1981)and James (1983) have found that many severely infected trees lack the crown symptoms thatare necessary for photo identification. Wallis and Bloomberg (1981) indicated that only onehalf of root diseased trees within and adjacent to disease centers could be determined fromabove ground symptoms. Wallis and Bloomberg’s work dealt with Phellinus weiril (Mum)Gilbertson, but similar findings are likely for A. ostoyae.Ground survey methods are probably most appropriate for assessing A. ostoyae extentand impact. The ground survey method developed for P. weirli by Bloomberg et a!. (1980),along with modifications for multi-disease recording and stratification by disease intensity(Bloomberg 1983), is applicable for surveys of A. ostoyae (Morrison et a!. 1991a). However,this method is difficult to apply in logged, burnt or open stands with diffuse diseasedistribution due to difficulty in locating boundaries of infection (Morrison et at. 1991a). Sincemany of the stands in the ICH biogeoclimatic zone are logged, burned or contain diffuse A.ostoyae infections, the applicability of the Bloomberg method in this zone is questionable.Another ground survey method, the Pixel survey (Merler and Norris, B.C. Ministry ofForests, unpublished report), uses a series of systematically located transect lines runningperpendicular to a base line (Morrison et a!. 1991a). Pixels are plots that vary in sizedepending on the age and type of stand to be surveyed. Pixels are contiguous on both sides ofthe transect lines. Pixels are identified as being either diseased or disease free based onevidence of signs and symptoms of disease within their boundaries. Pixel surveys can provideinformation on distribution of A. ostoyae infected areas. A modification to the pixel surveywhich involves weighting trees based on dbh and infection status has recently been developed.15The modification renders a better estimate of disease intensity (Merler, B.C. Ministry ofForests, unpub.)Armillaria ostoyae is known to occur in a large proportion of the ICH zone (Morrisonet al. 1991a). Based on field observations in two study areas located in the ICH zone aroundShuswap Lake, the extent of A. ostoyae infections in mature and immature stands within thiszone is indeed great. Due to the extremely wide distribution of A. ostoyae in the ICH zone,the value of any detection survey made using any of the various methods, is questionable. Ifthe disease is present almost everywhere, why survey to detect where it is? It is clear that theamount of disease detected using any of the survey methods depends on the amount of timeand effort spent examining.1.4 SummaryArmillaria ostoyae is one of the most serious forest pathogens in the southern interiorof British Columbia. Several features of the Armillaria genus set it apart form other membersof the Agaricales, including the production of rhizomorphs, diploid mycelium and the presenceof both a parasitic and a saprophytic phase. The ability of A. ostoyae to colonize a host as aparasite and then survive in that host saprophytically for extended periods of time poses thegreatest challenge for forest managers in their attempts to manage for the disease. Theprogress of A. ostoyae infections depends on the inoculum potential of the pathogen.Inoculum potential is defined as “the energy of growth that is available to the pathogen toinfect a new host” (Garrett 1960). Inoculum potential depends on host species, food base size,degree of colonization by the disease, distance from an inoculum source, time sincecolonization and activity of other fungi and insects. The inoculum potential of a site can beinfluenced to a large extent by forestry operations and by fire. The amount and distribution ofinoculum on a site dictates the spread of the disease. Infection of new hosts takes place eitherby rhizomorphs penetrating the bark of by root contact with an infected host.16Trees infected by A. ostoyae exhibit symptoms and specific signs that may be used toidentify the disease. Symptoms include reductions in shoot growth, changes in foliarcharacteristics, stress induced reproduction, basal resinosis, wood decay and death. Signs ofA. ostoyae include mycelial fans, rhizomorphs and basidiomes. The symptoms and signs of A.ostoyae along with the spatial distribution of infected trees within stands are collectivelyreferred to as the disease’s signature. The signature of A. ostoyae is used in surveys designedto estimate the extent and impact of the disease.Most of the research regarding A. ostoyae in unmanaged forests in British Columbiahas been concerned with the disease at only one point in the life of the stands. There has beenvery little documentation on the impacts of A. ostoyae in the transition from unmanagedmature stands to managed plantations. As a result, it is not known whether the plantationsestablished today, in areas affected by A. ostoyae, will be able to produce timber volumessimilar to those that were harvested previously. With this information the decision of whetheror not to take actions to manage for the disease would be made much clearer.Armillaria ostoyae is known to occur in a high proportion of the ICH zone in thesouthern interior of British Columbia. It is not known, however, which areas within this zoneare most at risk to the impacts of A. ostoyae. With these two pieces of information thedecision of whether or not to take actions to control the disease on a specific site would bemade much easier.17CHAPTER 2 ARMILLARIA OSTOYAE DEVELOPMENT IN PLANTATIONS,FOLLOWING THE HARVEST OF INFECTED STANDS2.1 IntroductionA considerable amount of the research regarding A. ostoyae in western north Americahas dealt with the disease in natural forests. James et a!. (1984) estimated the losses in forestproductivity in the northern Rocky Mountain National Forests in northern Idaho andnorthwestern Montana resulting from A. mellea and P. weirii infections. Their study providedan estimate of current losses to these diseases at one point. They estimated that root diseasesaccounted for 35 % of the annual tree mortality on a three million hectare forest. The volumeloss associated with the diseases was 834 000 m3. There have been several similar studiesconducted in the northwest United States (e.g., Filip and Goheen 1982, Stewart et a!. 1982,Williams and Leaphart 1978, Filip 1977) that were designed to estimate either the areainfected by, or the volume losses attributable to A. ostoyae in mature stands. Filip andGoheen (1982) estimated that over the past 20 years, 21.6% of the merchantable volume in aseverely infected stand in central Oregon was lost to root disease. These loss estimates havehelped forest managers select stands for treatment and have aided forest planning (James et a!.1984). However, more complete information of productivity losses, is still required.Growth reduction in trees infected with A. ostoyae also contributes to the loss in overallforest productivity. Bloomberg and Morrison (1989) addressed this problem in their study ofthe relationship of growth reduction in Douglas-fir to infection by A. ostoyae in southeasternBritish Columbia. They found that growth reductions varied among the infected stands.Bloomberg and Morrison (1989) surmised that the differences in growth reduction wereprobably due to differences in initial disease levels and differences in site and stand conditions.They concluded that individual losses in tree growth were substantial and cumulative over longperiods of time, and that growth reductions were due to the destruction of the host rootsystem.18The research in A. ostoyae mentioned thus far deals with losses through mortality andgrowth reduction, in mature natural stands. There is a great need for a better understanding ofthe behavior of A. ostoyae in these mature stands once the areas are harvested and replanted.There is very little information describing how the disease develops in successive crops plantedon the same site, except that A. ostoyae persists in subsequent rotations in forest plantations(Hood et al. 1991). The extent of disease spread and the future losses that will occur in theseinfected areas is not known. A stump removal trial located near Sicimikin, B.C. (Morrison eta!. 1988) tested the differences in susceptibility between species and the efficacy of stumpremoval as a means of reducing the inoculum of A. ostoyae and P. weirii. The trial wasoriginally established to investigate the effectiveness of inoculum control treatments for P.weirli. Hence, the amount of P. weirli inoculum at the outset of the experiment wasdetermined. At that time, A. ostoyae was only considered to be a secondary pathogen of treesattacked by P. weirii. This may account for the absence of records of A. ostoyae incidence inthe original stand (Morrison et a!. 1988).The Skimikin trial provides insights into the differences in tree species susceptibility toA. ostoyae from plantation establishment up to age 25. The trial provides an example of whathappens when a root-disease-infested stand is harvested and replanted with susceptible hosts.It clearly demonstrates the value of removing large sources of inoculum in order to improveplantation survival. However, the question still remains, what was the extent of A. ostoyaeinfection in the previous stand?With the exception of Morrison and Pellow (1993) very little research has attempted todirectly tie the extent of A. ostoyae infections in mature stands to the impact of the disease inplantations established on the same sites. Little is known about how individual infectioncenters behave in older plantations (Hood et al. 1991). The reason for this lack of research isdue in large part to the length of time required to track the progress of the disease into olderplantations. Permanent sample plots established in infected plantations would be ideal;however, they are expensive and they require a long period of time before any meaningfulresults are available. Preferably, permanent sample plots would be established prior to harvest19so that the initial levels of disease in the mature stand could be related to future diseaseexpression in the plantations. Such permanent sample plot installations would requireconsiderable foresight.One objective of this research was to develop a better understanding of A. ostoyaedevelopment in plantations following the harvest of infected stands in the ICH biogeoclimaticzone. In order to achieve this objective, without the use of permanent sample plots, a meansof estimating infection levels prior to harvest in older plantations was required. The fact thatmycelial fans of A. ostoyae leave impressions in the resin and bark of conifers for severalyears after the mycelium itself has disappeared provided that means.Mycelial fan impressions have been found on stumps that were created over 30 yearspreviously, according to B.C. Ministry of Forests records. Within these areas mycelial fanimpressions were also found on snag stumps (I. e, stumps that were not a result of logging) thatwere probably as old as 50 years or more, based on visual comparisons with the loggedstumps. The mycelial fan impressions on stumps provide evidence of past A. ostoyaeinfections which may be used to determine the levels of infection in former stands prior toharvest. In order to make this determination, it is clear that the mycelial fan impressions onstumps must provide evidence as to when the fan grew to the root collar. Armillaria ostoyaecolonizes newly created stumps in three ways: by rapid extension from pre-existing lesions(Kile 1980), by invasion from an epiphytic position on the roots, or by invasion from outsideby newly arrived rhizomorphs (Redfern and Filip 1991). In order to determine pre-harvest A.ostoyae infection levels, it is essential to differentiate between the mycelial fan impressions leftby the fungus post-harvest from those impressions that were visible at the root collar in thestand prior to disturbance.Based on field observations in stumps cut the previous year, it was discovered thatthere were subtle differences in A. ostoyae mycelial fan morphology. Mycelial fans foundbeneath resinous bark at or near the root collar were deeply ridged, short (each fan less than5cm long), contorted and associated with resin. There was an abrupt transition at the edge ofthese short, contorted fans where the morphology was altered. These new mycelial fans were20less deeply ridged, expansive (often longer than 20cm), and free of resin. As the largermycelial fans were free of resin and appeared to develop from mycelial fans that containedresin, it was believed that the larger mycelial fans had been formed following the cutting of thetree. One principle response of conifer hosts to attack by A. ostoyae is the production of resin(Morrison et a!. 1991b). In order for the host tree to produce resin it must have been alive.The differences in the size and shape of the mycelial fans could also be explained bythe lack of host resistance in the cut stumps. In live hosts, mycelial fans are typically deeplyridged, short, and contorted while those found in dead snags are relatively smooth and longextending over 20cm. In the absence of host resistance the mycelial fans of A. ostoyae spreadrapidly, creating large sheets of mycelium.Armillaria ostoyae is rarely capable of infecting dead trees or stumps because it is apoor competitor in relation to strict saprophytes. The stumps with infections classified in thisstudy as post-harvest were in fact infected prior to harvest, but the evidence of infection wouldnot have been visible above ground in the undisturbed stand. The stumps with “post-harvest”infections, in effect, provide an estimate of the number of A. ostoyae infections that would bemissed in a typical root disease survey. Those infections classified as “pre-harvest” wouldhave shown signs of infection at the root collar. These signs would have been in the form ofmycelial fans beneath the bark and perhaps basal resinosis.Mycelial fan impressions of both forms of A. ostoyae mycelium were found in the barkof old stumps within all of the study areas. Thus, it was possible to identify which stumps hadshown above ground signs of infection prior to harvest and those which did not exhibit thisevidence until after harvest. This evidence allowed for an estimate of the number, proportionand basal area of stumps with root collar lesions caused by past A. ostoyae infections in thefour plantations used in this study. This evidence was used to determine if future losses inplantations due to this disease could be predicted on the basis of root disease surveys both priorto and following harvest. It was also used to determine which measurement of stumpinfections, the number, proportion or basal area, was the best for predicting mortality in futureplantations. The differing ages of the plantations in this study allowed for an examination of21the impacts of A. ostoyae in plantations over time. The impacts of A. ostoyae includedchanges in species composition among the stands. Differences in plantation histories allowedfor an examination of the effects of brushing treatments on the behavior of A. ostoyae. Theuse of two different plot sizes in the study allowed for an examination of the relationshipbetween infected stumps and the probability of regeneration close to that stump being infected.2.2 Methods2.2.1 Study AreasFour plantations were selected as study areas. All four plantations were located withinthe Interior Cedar Hemlock biogeoclimatic zone, in the Shuswap moist warm Interior CedarHemlock variant (ICHmw2) (Lloyd et al. 1990). The study areas lie on the east slopes ofLarch Hills, close to Salmon Arm, British Columbia. The plantations chosen were selectedusing the following criteria: suitable site preparation (broadcast burned rather thanmechanical), clearly defined date of establishment, reasonable proximity to each other, andease of access. The information used to select the four plantations was obtained from a B.C.Ministry of Forests Forest Cover Map (82L.075), scale 1:20 000, produced by the InventoryBranch of the Ministry of Forests. The history of each of the plantations was described on theForest Cover Map and Forest Service records (Table 2).Once the four plantations were selected, a preliminary walk through was conducted todetermine if A. ostoyae was present on the stumps. A test of the proposed sampling methodswas also completed.22TABLE 2. Forest management history records for the four plantationsPLANTATION 1969a 1969b 1980 1984HARVESTED 1968 1965 1974 1979BURNED 1968 1967? 1976 1980PLANTED 1969 1969 1980 1984SPECIES PLANTED Fdi Fdi P1 Fdi/P1TREES/HA 1000 1300 1000 1200BRUSHING - 1984/1986 - 19872.2.2 Sampling DesignFifty, 8m-radius-fixed-area plots (0.O2ha) were established on a 100 x lOOm gridwithin three of the study areas. The fifty 0.O2ha plots resulted in a total area sampled in eachplantation of lha. The 1980 plantation was sampled with 48 plots as its size and shape did notallow for 50 plots in the grid. The strip lines were laid out on cardinal bearings. Plots werelocated by tight chaining using a 50m nylon chain and a silva ranger compass.2.2.3 Sampling ProcedureOnce the plot center was established, a 30m Eslon tape was used to locate thecircumference of the plot using an 8m radius. The plot perimeter was flagged with florescenttape. The plot was then divided into four sectors. Within each sector, all stumps over 10cmdiameter at root collar (DRC) were examined. Stumps species, DRC, origin and disease statuswas recorded accordingly. Species identification was based on bark and wood characteristics.Due to the age of the stumps, some of which were cut in 1965, species identification wassometimes difficult. Stump origin differentiated snags from cut stumps. Disease status wasdetermined by examining under the bark from the cut surface to the root collar for evidence ofmycelial fan impressions. Due to the age of the stumps and the fact that all four plantationswere burned for site preparation, many stumps did not have completely intact bark. In suchcases, careful examinations under the bark in the crotches of roots at the root collar wereperformed. Root excavations were not performed in this study. The broadcast burn site23preparation method had exposed some roots below the root collar by burning off the dufflayer. Much of the below ground portions of these stumps were too badly decomposed toyield any evidence of A. ostoyae mycelial fans. Stumps were identified as belonging to one ofthe following four classes:CLEAN - stumps that still had most of their bark intact and exhibitedno signs of A. ostoyae mycelial fan impressions;ARMILLARIA- stumps where portions of the bark exhibited evidence ofBEFORE mycelial fan impressions that were short, deeply etched,contorted and often associated with resin, and stumps withcallused lesions surrounding bark with any mycelial fanimpressions;ARMILLARIA - stumps on which portions of the bark exhibited evidence ofAFTER mycelial fan impressions that were long, lightly etched andexpansive with no evidence of resin and no evidence of A.ostoyae before, and stumps with extensive colonization bydichotomously branched rhizomorphs beneath the bark;NO EVIDENCE - stumps which no longer exhibited enough evidence to beclearly placed in any of the other categories.It is possible that these classification methods were biased. A stump could bepositively identified as infected based on a small portion of the inner bark exhibiting mycelialfan impressions or on the presence of rhizomorphs. In order for a stump to be classified asclean, a large proportion of the inner bark yielding no signs of past infection was required. Ifonly a small proportion of the inner bark was intact and exhibited no signs of disease it wasnot clear that the stump had not been infected. The missing portions of inner bark may havebeen infected yet there was no way to determine if this was the case. Such stumps were placedin the no-evidence class. Thus, it was less likely for a stump to be classified as clean than tobe classified as infected since decay and weathering over time lead to the sloughing of stumpbark.Following stump examinations, the species, height and DRC of all regeneration withineach sector was recorded along with any evidence of disease. Height and diametermeasurements were visually estimated to the nearest meter and centimeter respectively, and24periodically verified. Regeneration infected with A. ostoyae was classified as either dead ordying. Armillaria ostoyae infections were confirmed by the presence of mycelial fans ormycelial fan impressions beneath the bark at the root collar. Evidence of other disease andinsect damage was recorded in comments. In the plantations that had received brushingtreatments (1969b and 1984) the brushed stumps were also examined. The DRC and diseasestatus was recorded. Due to excessive stocking levels and resprouting brush competition,some of the brushed stumps could have been missed.A small survey was also carried out in the mature stands surrounding each of theplantations. The purpose of this survey was to determine whether the pre-harvest levels ofinfection identified in the stumps was similar to that found in the surrounding mature stands.This survey acted as a check to back up the claim that past A. ostoyae infections could beidentified as being either pre- or post-harvest based on the morphology of mycelial fanimpressions.Ten fixed area plots of O.O2ha (8m radius) were established in three strips more thanlOOm outside the cutblock boundaries of three of the plantations. Each tree within the plotwas closely examined for any evidence of A. ostoyae infection. The humus layer was removedup to 30cm below the root collar in order to simulate the effects of the broadcast burntreatments that were performed on the plantations. As mentioned earlier, the stumps in theplantations often had portions of their roots below the root collar exposed, most likely due tofire. Bark was removed from much of the roots and root collar using an axe. This wasnecessary since the identification of A. ostoyae on the stumps in the plantations often requiredremoval of all the bark in order to positively identify the disease.252.2.4 Data AnalysisThe data was entered into Quatro Pro (Version 3.01. 1991), a spreadsheet program,and then transferred into Systat (Version 5.03. 1991) for analysis. Each plantation wassummarized independently. Within plantations, stump data and regeneration data were firstsummarized independently and later combined. All analyses were conducted using oc=O.05unless stated otherwise in the results section.StumpsThe original stump data consisted of species, DRC, stump origin and disease status.Basal area in square meters per stump was calculated as (DRC/200)2*ir.Stump data were summarized by species, disease status and basal area, such that foreach plot the following information was determined:- total basal area by species (TOTALBA)- basal area infected pre-harvest (BABEFOR) by species1- basal area infected post-harvest (BAAFTR) by species2- clean basal area/species (CLEANBA)- no-evidence basal area/species (NOEVBA)- total number of stumps/species- number of pre-harvest infected stumps/species (STMPBEFR)- number of post-harvest infected stumps/species (STMPAFTR)- number of clean stumps/species (CLEAN)- number of no-evidence stumps/species (NOEVID)- proportion of all stumps infected pre-harvest (ARMSEVB)3- proportion of all stumps infected post-harvest (ARMSEVA).1snag stumps showing evidence of past A. ostoyae infections were classified as being infected pre-harvest andcombined with the other pre-harvest infected stumps.recorded as infected post-harvest included all stumps ex1ibiting evidence of A. ostoyae infections sincestumps infected pre-harvest would remain infected post-harvest.31,oth of these proportion measures included the no-evidence stumps in the total number of stumps.Chi-square tests were performed on stump data for the three major species (westernredcedar, Douglas-fir and western hemlock) to determine if there were significant differencesin their relative frequencies of infection (I. e, were the ratios of infected stumps versus clean26stumps similar between the three species). This test revealed no significant differences. Basedon these results, the basal area, stump frequencies, and infection proportions for all specieswere combined to arrive at single values for each of these variables for each plot. This wasnecessary in order to make direct comparisons between the infection status of stumps andregeneration within the same plots. This aspect will be described in detail later.The stump data were also summarized by species, origin, plantation and diameter class.The following relationships were examined:- stump frequency by disease category among species- stump frequency by disease category between mature andimmature stands- stump frequency by disease category among species andbetween stump origin classes- stump frequency by origin among plantations- stump frequency by species and diameter class- stump frequency by disease category among plantations- stump frequency by species and plantation- basal area by disease category among plantationsStump Frequency Distribution by Species and Disease StatusDue to the very low numbers of stumps for all other species, only western redcedar,Douglas-fir and western hemlock stumps were analyzed for differences in stump distributionsbetween the disease identification classes. Chi-square tests were used to test for differences inthe frequency of A. ostoyae infection both before and after harvest and for differences in thepreservation of evidence among western redcedar, western hemlock, and Douglas-fir. Thefirst Chi-square test included all three species and indicated that there were significantdifferences among the three species in their relative proportions of stumps in each of the fourinfection status classes. Further Chi-square tests were performed on the distribution of stumpsby species between each infection status class individually. This was accomplished bycomparing the proportion of stumps for each species in each of the four infection classes.27Rates of Infection in Stumps Compared to those in the Surrounding Mature StandsThe infection rates found in the stumps of all four plantations combined were comparedto those found in the surrounding mature stands. A Chi-square test was used to test fordifferences between the two areas. All tree and stump species were combined for the analysis.Stump Species and Disease Status by Stump OriginThe majority of both snag and cut stumps belonged to one of three species: westernredcedar, Douglas-fir, and western hemlock. A Chi-square test was used to test fordifferences among these three species in their relative proportions of snag stumps to cutstumps. All three species were compared in a 3 by 2 contingency table. Further pairwisecomparisons using 2 by 2 contingency table tests between the species were used to determinewhich of the species were significantly different from each other.Stump Origin by PlantationA 4 by 2 contingency table Chi-square test was used to test for differences among thefour plantations in the proportion of snag stumps out of the total stump count. Pairwisecomparisons between the plantations were performed to determine which of the plantations hadsignificantly different proportions of snag stumps.Stump Frequency Distribution by Species and Diameter ClassThe diameter class distributions of the three major species were compared. The overalldistribution of stumps among the 10 diameter classes was compared between combinations ofspecies using 10 by 2 contingency tables. Differences among the three species were testedwith Chi-square tests at each of the 10 diameter classes using pairwise comparisons between28the proportion of stumps in each diameter class for each species in all diameter classescombined.Stump Frequency Distribution by Disease status and PlantationThe proportions of stumps in each disease class out of the total number of stumps(excluding the no-evidence stumps) were compared among the four plantations. Including theno-evidence stumps would reduce the proportions of infected and clean stumps. The truestatus of the no-evidence stumps could not be determined. An equal distribution of thesestumps between the other three classes, infected pre-harvest, infected post-harvest, and clean,represents a best guess. The lack of evidence in stumps was most often due to ants consumingthe inner bark. A series of Chi-square tests were used to test for differences in the proportionof stumps in each of the disease classes among the four plantations. The proportion of stumpsthat were clean, infected pre-harvest, and infected post-harvest out of the total number ofstumps were compared among the four plantations. Plantations were compared using pairwiseChi-square tests (1 degree of freedom).Stump Frequency Distribution by Species and PlantationThe relative proportions of the three major species were compared among the fourplantations using a series of Chi-square tests. The first Chi-square test indicated a significantdifference in species composition among the plantations. The proportions of stumps in each ofthe species out of the total for each plantation were then compared in pair-wise combinationsusing 2 by 2 contingency tables (1 degree of freedom).29Stump Basal Area Distribution by Disease Status and PlantationA one-way ANOVA was used to test for differences in the mean plot basal area ofstumps infected before and after harvest and those stumps classified as no-evidence betweenthe four plantations. A Tukey multiple range test was then used to determine which of theplantations were significantly different from each other. To express the basal area values asper hectare units the mean basal areas/plot were multiplied by 50. Basal area in this studyrefers to the cross sectional area of the stump at root collar and not at breast height.RegenerationRegeneration data for each plot were summarized by species and disease status suchthat for each plot the following information was determined:- total number of trees (TOTREES)- total number of conifers/ha (CONSTEMS)- number of conifers infected with A. ostoyae (TREINFCT)-and, proportion of conifers infected (PROINFCT)The proportion of conifers infected was determined by dividing the total number ofconifer trees infected, all species and sizes combined, by the total number of conifers. Plantedtrees could not always be distinguished from the natural regeneration so the two could not beseparated in the analysis.Regeneration data were also summarized by species and 5cm diameter classes in orderto determine relative rates of infection among the species by diameter class. The followingrelationships concerning the regeneration in the four plantations were examined:30- species composition among plantations- relative rates A. ostoyae incidence among conifer species- regeneration species distribution by disease status and diameter class- relative rates of infection among species in smallest diameter class (0-2.5cm)- diameter class distribution for each plantation- relative rates of regeneration infection among plantations- distribution of infected regeneration within plantations by plotSpecies Composition Among PlantationsEight conifer species were present in each of the four plantations. The distributions ofthese species in relation to plantation, diameter class, and disease status were compared anddescribed. Deciduous species were not included in the analyses although the data werecollected. Two of the plantations had received brushing treatments and therefore fewdeciduous stems remained in these areas.Comparison of the Relative Rates of Armillaria ostovae Infection Incidence among ConiferSpeciesThe relative rates of A. ostoyae infection incidence among conifer species werecompared among the eight species found throughout the four plantations. The average rate ofinfection and the species composition varied among plantations. Thus, averaging theproportion of trees infected by species across all four plantations would confound the rates ofinfection incidence by possible differences in disease severity among sites. A standardmeasure of disease severity for each site was required so that comparisons among species couldbe made given the variability among the plantation sites. The proportion of Douglas-firinfected on a site was chosen as a standard measure of disease incidence, since this species wasrelatively abundant and generally accepted as susceptible to A. ostoyae infection. The rate ofinfection of each species was then expressed as a ratio to the rate of infection in Douglas-firfor each plantation. These ratios of infection were then tested using a Chi-square test todetermine whether or not they were constant across the four plantations. The expected valuesrequired for the Chi-square test were calculated using the following formula:31Expected value = SI1 * * KWhere: SI1 = the proportion of Douglas-fir infected in plantation (i)n1 = the number of species (x) in plantation (i)K = total number of species (x) trees infected* n)The K-value was a constant for each species that was used to test the ratios amongplantations. A Chi-square test was used to determine if there were significant differencesamong plantations in the proportion of trees infected for a given species as compared to theproportion of Douglas-fir infected. The null hypothesis for this test was that there was nosignificant difference between the proportion of species (x) infected and the proportion ofDouglas-fir infected over the four plantations. In other words, on those plantations whereDouglas-fir was heavily infected, species (x) would be expected to be heavily infected as well.A significant Chi-square result would indicate that the proportion of species (x) infected didnot vary in conjunction with that of Douglas-fir. Such a result would indicate that the relativedifferences in disease incidence among species was perhaps more dependent on site factorsthan on inherent differences among species. The K-value was a measure of relative diseaseincidence among species using Douglas-fir as a baseline. A conifer species with an K-valuegreater than one would, therefore, have a relatively greater proportion of trees infected on asite than Douglas-fir. The reverse could be said for an K-value less than one. The relativerates of infection were compared between lodgepole pine and Douglas-fir (Table 3).TABLE 3. Relative rates of infection compared between lodgepole pine and Douglas-fir regenerationSite Fd SI Total P1 Infected P1 K Expected CM-square1969a 0.04394 7 0 1.0484 0.3225 -1969b 0.17995 18 1 1.0484 3.3961 1.9871980 0.12618 1021 134 1.0484 135.071 0.00851984 0.07022 329 28 1.0484 24.213 .593Total 1375 163 163.0 2.58932If the expected number of a given species was less than one, that expected value wasadded to the expected value for the plantation with the largest relative difference betweenexpected and actual values. In the example above, the expected number of lodgepole pinetrees infected was 0.3225 in the 1969a plantation. This expected value was added to the valuefor the 1969b plantation since the relative difference between expected and actual values wasgreatest in this plantation. These methods ensured a most rigorous test of the null hypothesis.Only the greatest differences in the relative proportions of trees infected between the givenspecies and the Douglas-fir standard would be significant among plantations.Once the K-values had been determined, it was necessary to determine which of thevalues were significantly different from each other. The proportion of trees infected werecompared among all eight conifer species among the four plantations to determine whichspecies were significantly different from each other. Every possible species comparison wasmade for a total of 28 tests. The proportion of Douglas-fir and lodgepole pine trees that wereinfected are compared (Table 4).TABLE 4. Observed and expected numbers of infected lodgepole pine and Douglas-fir regeneration within the four plantationsSITE Lodgepole Pine Douglas-Fir % INfECTED# TREES INFECTED # TREES INFECTED OVERALLobs. exp. obs. exp.1969a 7 0 0.305 751 33 32.695 0.04351969b 18 1 3.188 778 140 137.81 0.17711980 1021 134 132.775 317 40 41.224 0.13001984 329 28 24.169 1182 83 86.831 0.0735In this example the null hypothesis was that in each plantation the percent infected ofDouglas-fir and lodgepole pine were the same. The resulting Chi-square = 2.260, thereforethe null hypothesis cannot be rejected. There is no significant difference in the proportion oftrees infected between lodgepole pine and Douglas-fir.33The expected values were determined by calculating the overall percent infected for bothspecies combined for each plantation then multiplying the number of trees in each species bythat overall percentage. Each plantation had a separate overall percentage of trees infected.The expected values were then compared to the observed values in a Chi-square test. Thedegrees of freedom for the Chi-square test was the number of terms compared, minus thenumber of overall percentages. In the example above, the degrees of freedom was 3 (7 termsminus 4 overall percents infected). If the calculated expected value was less than 1, that valuewas added to the expected value that differed the most from the observed value for a givenplantation and species.Regeneration Species Distribution by Disease Status and Diameter ClassThe frequency distributions of each of the eight species by 5cm diameter class werecompared among the regeneration species. The distribution of infected regeneration among thesame five, 5cm diameter classes was also examined. The distribution of diseased stemsthroughout the range of diameter classes were compared among the four major species,western redcedar, Douglas-fir, western hemlock and lodgepole pine.Spatial Distribution of Infected Regeneration within PlantationsThe spatial distribution of infected trees was examined within each of the plantations.The total number plots from each plantation were divided into five classes based on thenumber of infected trees per plot. The classes ranged from 0 infected trees/plot to greater than6 infected trees/plot.Stumps and RegenerationTo analyze the relationship between past and present levels of A. ostoyae on a site, itwas necessary to determine individual plot values for both stumps and regeneration. The34individual stump plot value could then be compared to the individual regeneration plot valuefor that plot in a correlation analysis. To arrive at a single plot value for stumps, all species ofstumps were combined.For each plot, the following stump measures were determined:(1) number of stumps infected pre-harvest (STMPBEFR)(2) number of stumps infected post-harvest (STMPAFTR)(3) proportion of stumps infected pre-harvest (ARMSEVB)(4) proportion of stumps infected port-harvest (ARMSEVA)(5) basal area of stumps infected pre-harvest (BABEFOR)(6) basal area of stumps infected post-harvest (BAAFTER)The relationships between the number of conifers infected (TREINFCT), the proportionof conifers infected (PROINFCT), and the number of disease free conifer stems/ha(CONSTEMS), and the variables listed above, were tested using correlation analyses. Thecorrelations were performed on the data from all four plantations combined and from eachplantation individually. These tests were designed to determine if the losses due to A. ostoyaein plantations were related to infection levels in the previous stand.Comparison of Measurements of Stump InfectionCorrelation analyses were used to determine which measure of stump infection out ofthe six listed above was most closely related to regeneration mortality if such a relationship didexist. All six measures of stump infection were also compared simultaneously in a stepwisemultiple linear regression. This method was used to determine which of the six factorsaccounted for the greatest degree of variation in the number of conifers infected, theproportion of regeneration infected, and the number of disease free stems/ha.35Effect of Plot Size on the Strength of Relationships Between Infected Stumps and InfectedRegenerationThe effect of plot size on the relationships between infected stumps and infectedregeneration was analyzed. The correlations between the total number of infected stumps andthe number of infected regeneration on full sized plots in all four plantations were determined.This was repeated on a sector (quarter plot) basis in the 1969b and 1984 plantations. The firstsector in each of the plots was selected for analysis. A single sector was chosen so that thesample size for both the full size plots and the sectors would be the same (50).The 1969b plantation was chosen since the relationships between infected stumps andvisibly infected regeneration were strongest in this plantation compared to the other three. The1984 plantation was chosen because it was the youngest of the four. The probability of stumpsbeing closely associated with dead or dying regeneration would be highest in the youngestplantation. As plantations age, the infection of new hosts by A. ostoyae depends more on rootcontacts between infected and healthy trees than on contacts between trees and infected stumps(Morrison pers. comm.).Analyses were run first with all plots and all number 1 sectors within the 1969b and the1984 plantations. Following this, the same plantations were analyzed using only those plotsand sectors that contained any infected regeneration. An analysis using only those samplingunits containing infected trees would conceivably reduce the variation in the relationship. Ifthere was a relationship between infected stumps and infected trees, then removing thosesampling units that contained no infected trees from the analysis would place a higheremphasis on those that did contain such trees.362.3 Results2.3.1 Stump ResultsStump Frequency Distribution by Species and Disease StatusThe vast majority of stumps (94%) throughout all four plantations belonged to one ofthree species: western redcedar (Cw), Douglas-fir (Fd), or western hemlock (Hw). Theremaining 6% of the stumps consisted of sub-alpine fir (Bl), paper birch (Ep), western larch(Lw), lodgepole pine (P1), western white pine (Pw) and Engelmann spruce (Se). Westernwhite pine actually made up 2.6%, and Engelmann spruce contributed 1.1 % of the totalstumps. Table 5 describes the distribution of stumps among the infection status classes byspecies for all four plantations. The percents of stumps infected in Table 5 were calculated bydividing the number of stumps in each infection status class by the total number of stumps.An alternative method involved percent calculations based on only those stumps that could bepositively identified as being clean or infected (Table 6).TABLE 5. Percentage of total stumps by Armillaria ostoyae infection status in eachspecies for all four plantationsCw Hw Fd Pw Se Er Lw P1 Bi MEANCLEAN 21.0 27.6 25.4 4.9 7.4 26.1 13.3 20.0 20.0 23.6Arm.BEFORE 35.3 32.9 36.6 59.0 37.0 13.0 20.0 40.0 60.0 35.3Arm.AFTER 15.1 30.0 26.9 13.1 37.0 17.4 46.6 0.0 20.0 23.3NOEVIDENCE 28.6 9.4 11.1 23.0 18.5 43.5 20.0 40.0 0.0 17.7TOTAL STUMPS 872 735 642 61 27 23 15 5 5 2385The results (Table 7) demonstrate that there was no significant difference in theproportion of stumps infected prior to harvest among the three major species. Infection classesthat did indicate significant differences among species were further tested using pairwise37combinations and Chi-square tests to determine which species were significantly different fromeach other.TABLE 6. Percentage of total stumps positively identified by infection statusin each species for all four plantationsCw 11w Fd Pw Se Ep Lw P1 BI MEANCLEAN29.4 30.5 28.5 6.4 9.1 46.2 16.7 33.3 20.0 28.7ARM.BEFORE 49.4 36.3 41.2 76.6 45.5 23.1 25.0 66.7 60.0 42.9ARM.AFTER 1.2 33.2 30.3 17.0 45.4 30.8 58.3 0.0 20.0 28.3TOTAL STUMPS(identifiable) 623 666 571 47 22 13 12 3 5 1962TABLE 7. Comparison of the proportion of stumps infected pre- and post-harvest,clean and no-evidence among the three maj or species in all fourplantations combined (proportions expressed in percent of total in thatspecies)Cw Fd 11w TOTAL CHI-SQUARECLEAN 21.0 25.4 27.6 549 10. 148(1) (1) (2)PRE-HARVEST 35.3 36.6 32.9 785 2.256(1) (1) (1)POST-HARVEST 15.1 26.9 30.1 526 56.025(1) (2) (2)COMBINED 50.5 63.6 63.0 1311 36.05INFECTIONS (1) (2) (2)NO EVIDENCE 28.6 11.1 9.4 389 126.101(2) (1) (1)TOTAL 872 642 735 2249Percentages with the same number in brackets below them are not significantly different from the others in thatrow based on a critical Chi-square value of 3.841 (1 degree of freedom) and pairwise tests. Critical Chi-squarevalue for the entire row comparisons is 5.991 (2 degrees of freedom).The proportion of stumps infected post-harvest could be viewed as a measure of thosestumps that did not exhibit above ground symptoms of infection at the time of harvest. SinceA. ostoyae can not compete well with other saprophytes, it is very likely that those stumps38with infections classified as post-harvest were already infected at the time of harvest. Theresults (Table 7) indicate that 15.1% (132 out of 872) of western redcedar stumps wereinfected prior to harvest without showing above ground evidence. The other two majorspecies had significantly more stumps in this category (26.9% and 30.1% of Douglas-fir andwestern hemlock stumps respectively, had infections prior to harvest without exhibiting anyabove ground symptoms). These results have important implications for pre-harvest rootdisease surveys. If the pre-harvest estimates of the extent of root disease infection in a standmay be off by as much as 30%, the value of a survey is questionable.Rates of Infection in Stumps Compared to those in the Surrounding Mature StandsThe proportion of stumps infected with A. ostoyae in the plantations was compared tothe proportion of mature trees infected in the stands surrounding the cut-over areas. Theresults of this test (Table 8) showed that there were no significant differences between thestumps and the mature stands in their respective proportions of infected pre-harvest and cleanstumps.TABLE 8. Comparison of the percent of stumps infected with Armillariaostoyae in all four plantations to that in the surroundingmature treesCLEAN INFECTED PRE-HARVEST TOTALALL STUMPS 40.7 59.3 1324(198 plots)MATURE TREES 41.4 58.6 232(10 plots)TOTAL 635 921 1556CM-square critical value= 3.841 CM-square obtained =0.035There is no significant difference in the proportion infected and clean. (cr=0.05)39The results of this comparison indicate that the methods used in this study todifferentiate between pre- and post-harvest infections were defendable. The proportion ofstumps classified as being infected prior to harvest in the plantations was almost identical tothe proportion of trees in the mature stands that exhibited signs of A. ostoyae infection. Thus,it is likely that the estimates of pre-harvest infections in the plantations were quiterepresentative of the disease conditions in the former stands.Stump Species and Disease Status by Stump OriginOf the 2385 stumps examined, 2057 (86%) were cut stumps and 328 (14%) were snagstumps. The majority of stumps belonged to one of three species, western redcedar, Douglas-fir, or western hemlock. The same general species composition was found when only snagstumps were considered. Of the 328 snag stumps tallied, 297 (90.5%) were either westernredcedar, Douglas-fir, or western hemlock (Table 9). A Chi-square test indicated that therewere significant differences between the species in the proportion of stumps in each originclass (Chi-square critical=5.991, Chi-square obtained =29.62). There was no significantdifference between western hemlock and western redcedar (Chi-square obtained = .4 177, Chisquare critical=3.841) in the proportional distribution between the two origin classes. Therewas a significant difference between Douglas-fir and the other two species. A significantlygreater proportion of total Douglas-fir stumps were snag stumps compared to western hemlock(Chi-square obtained =22.68) and western redcedar (Chi-square obtained= 10.84).There was a significantly greater proportion of snag stumps infected with A. ostoyaethan cut stumps. The reason for this seems obvious since a major cause of mortality in naturalundisturbed stands would be A. ostoyae.40TABLE 9. Percentage of total stumps in each of the three major species ineach origin classCw Fd Hw TOTALPERCENT CUT 88.8 80.7 89.8 1952(1) (2) (1)PERCENT SNAG 11.2 19.3 10.2 297(1) (2) (1)TOTAL 872 642 735 2249Those percents with the same number in brackets below them are not significantly different fromthe others in that row (a = 0.05).Stump Origin by PlantationA Chi-square test indicated that there were significant differences among the fourplantations in the proportion of snag stumps out of the total stump count (Chi-squarecritical=:7.815, CM-square obtained= 19.552). There were significantly fewer snag stumps inthe 1984 plantation compared to the other three. The proportion of snag stumps in the 1969a,1969b and 1980 plantations were not significantly different from each other with 18.3, 16.0and 13.9% respectively. Only 10.0% of the 1984 plantation stumps were snag stumps.Stump Frequency Distribution by Species and Diameter ClassWestern redcedar and western hemlock stumps did not differ significantly in theirrelative proportions in each of the ten 10cm diameter classes. The critical Chi-square valuewas 16.9 19 while the Chi-square obtained was 14.22. Douglas-fir had a significantly differentoverall distribution of stumps among the 10 classes compared to the other two species, westernredcedar (Chi-square 120.97) and western hemlock (Chi-square 116.94). Douglas-fir tendedto have fewer small stumps (10-20cm) and relatively more large stumps in the 50-60cmdiameter classes (Table 10).41TABLE 10. Percentage of total stumps of the three major species in eachof ten 10cm-diameter classes all plantations combinedCLASS Cw Fd Hw TOTAL10cm 9.6 1.6 9.7 164(1) (2) (1)20cm 24.4 12.3 23.5 463(1) (2) (1)30cm 24.5 21.5 24.0 526(1) (1) (1)40 cm 17.4 21.4 18.6 425(1) (1) (1)50 cm 9.7 18.7 13.3 302(1) (2) (1)60 cm 7.4 15.4 5.9 206(1) (2) (1)70 cm 3.8 5.0 3.5 91(1) (1) (1)80cm 2.0 1.9 1.2 38(1) (1) (1)90cm 0.6 1.7 0.3 18(1) (2) (1)100 cm+ 0.7 0.5 0.0 9(1) (1) (1)TOTAL 866 641 735 2242Those percentages with the same number in brackets below them are not significantlydifferent from the others in that row (a=0.05)Stump Frequency Distribution by Disease status and PlantationThe frequency distribution of stumps for the four plantations is shown in Figure 1.There were large differences in the number of stumps/ha among the four plantations. Thelowest level was found in the 1969a plantation with 420 stumps/ha. The highest level wasfound in the 1984 plantation with 860 stumps/ha. There were also differences in the amountof disease evidence in stumps among the four plantations. The greatest number of infected42stumps occurred in the 1984 plantation with 515/ha compared to only 216/ha in the 1969aplantation.I—zD00aSTUMP DISTRIBUTION WITHIN PLANTATIONSFREQUENCY BY DISEASE CLASSFIGURE 1. Stump distribution among plantations by disease statusThe proportion of stumps identified as belonging to one of the three disease classes (I. e,clean, infected pre-harvest and infected post-harvest) were compared among the fourplantations (Table 11). In this table the proportion of stumps infected post-harvest was not ameasure of the total number of stumps infected. Post-harvest infected stumps in this analysisincluded only those stumps which exhibited signs of being infected post-harvest and did notinclude the pre-harvest infected stumps.TOTAL STUMPSINFECTED PRE-HARV.—INFECTED POST-HARV.CLEANNO-EVIDENCEPLANTATION43TABLE 11. Percentage of stumps infected with Armillaria ostoyae out of the totalidentified stumps (excluding no-evidence stumps) for each plantation inpercentPLANTATIONDISEASE STATUS 1969a 1969b 1980 1984 TOTALCLEAN 41.3 29.6 16.9 29.5 564(3) (2) (1) (2)PRE-HARVEST 38.9 50.5 46.5 38.5 842: (1) (2) (2) (1)POST-HARVEST 19.8 19.9 36.6 32.1 556(1) (1) (2) (2)TOTAL 368 402 462 730 1962Percentages with the same number in brackets below them are not significantly different from the othersin that row (a =0.05)The 1969a plantation contained a significantly greater proportion of clean stumps thanthe other three sites. The high proportion of clean stumps in this plantation probably in partreflects a difference in measurement system. This plantation was the first of the four to beexamined. In the early stages of data collection those stumps which lacked enough evidence tomake positive identifications were classified as clean. This different measurement systemwould also account for the low proportion of no-evidence stumps in the 1969a plantation (Fig.1).There was a significantly greater proportion of stumps infected pre-harvest in the 1969bplantation compared to the 1969a plantation. The measurement error described above could bepartially responsible for this difference since the proportion of clean stumps in the 1969aplantation is possibly over-estimated. The two youngest plantations contained significantlymore post-harvest infected stumps than the two older plantations. The mycelial fanimpressions left by post-harvest infections are more subtle than those left by pre-harvestinfections. It is quite possible that the post-harvest A. ostoyae evidence on the older stumpswas too weathered to be identified. The 1980 plantation was the most heavily infected of thefour plantations when pre- and post-harvest infections were combined.44Stump Frequency Distribution by Species and PlantationThe relative proportions of the three major species of stumps were compared among thefour plantations (Table 12).TABLE 12. Species composition of the stumps in each of the four plantations inpercentSPECIESPLANTATION Cw Fd Hw TOTAL1969a 42.0 21.9 36.1 393(1) (2) (2)1969b 40.5 20.4 39.1 511(1) (2) (2)1980 37.3 4.7 58.1 558(1) (1) (3)1984 37.1 54.1 8.8 787(1) (3) (1)2249The two oldest plantations, 1969a and 1969b, had very similar species compositionsbased on stump evidence. The 1980 and 1984 plantations had very different speciescompositions from each other and from the two older plantations. The proportion of totalstumps made up by western redcedar was not significantly different among the fourplantations. The major differences in species composition between the 1980 and 1984plantations were due to the relative proportions of western hemlock and Douglas-fir. The1980 plantation contained many more hemlock than Douglas-fir stumps while the reverse wastrue for the 1984 plantation. These differences in species composition are not clearly relatedto any differences in the number of stumps within the four disease status categories of eachplantation. The low proportion of western hemlock in the 1984 plantation could help toexplain the relatively high number of no-evidence stumps within this plantation. One wouldexpect the youngest plantation to contain stumps with the clearest evidence. The 1980plantation had the greatest number of western hemlock stumps per hectare (337) and it also hadTOTAL 872 735Those percentages with the same number in brackets below them are not significantly different fromothers m that column (a=0.05 Chi-square critical 3.841)45the lowest number of clean stumps per hectare (78). All three species had higher proportionsof infected stumps in the 1980 plantation.Stump Basal Area Disthbution by Disease Status and PlantationAs previously mentioned, the greatest number of stumps/ha was found in the 1984plantation. However, the greatest stump basal area was found in the 1980 plantation (Table13). The stump basal area in the 1980 plantation was 90.0m2/ha. Of that, 46.83m2/haor52% was infected pre-harvest. The mean plot basal area of stumps infected pre-harvest in the1980 plantation was significantly greater than the plot means found in the other threeplantations. As a consequence, the average clean basal area/plot in the 1980 plantation wassignificantly less than that found in the other three plantations. The mean clean stump basalarea/plot and the mean pre-harvest infected stump basal area/plot did not differ significantlyamong the 1969a, 1969b and 1984 plantations. The mean post-harvest infected stump basalarea/plot was not significantly different among the four plantations. The mean no-evidencestump basal area/plot was similar in all but the 1969a plantation.TABLE 13. Mean stump basal area for each plantation by disease status (m2/ha)PLANTATIONDISEASE STATUS 1969a 1969b 1980 1984 TOTALCLEAN 26.13 12.61 5.93 16.67 61.34(2) (1) (1) (1)INFECTED 27.66 25.91 46.83 31.03 131.43PRE-HARVEST (1) (1) (2) (1)INFECTED 16.70 12.04 24.10 25.79 78.63POST-HARVEST (1) (1) (2) (2)NO-EVIDENCE 7.32 15.29 13.15 11.79 47.55(1) (2) (2) (2)TOTAL 76.17 65.95 90.00 85.30 317.42Those percentages with the same number in brackets below them do not have significantly different mean basalareas from the others in that row (a=0.05). (Tukey test)46Stump Results SummaryThe majority of the stumps in all four plantations were either western redcedar,Douglas-fir or western hemlock. The proportion of no-evidence western redcedar stumps wasgreater than that of Douglas-fir or western hemlock. There were also significantly fewer post-harvest infected western redcedar stumps.The proportion of snag stumps for each plantation also differed among plantations.The 1984 plantation had the lowest proportion of snag stumps. The other three plantationswere not significantly different in this respect. Douglas-fir accounted for 41.8% of all snagstumps, while western hemlock and western redcedar accounted for 25.3 and 33.0%respectively.There were differences in the number of stumps/ha among the four plantations. Thelowest number was found in the 1969a plantation with 420 stumps/ha total. The highest wasin the 1984 plantation with 860 stumps/ha. However, total basal area did not varyconsiderably among plantations.There were also differences in the amount of disease evidence in stumps among thefour plantations. The greatest number of infected stumps occurred in the 1984 plantation with515/ha compared to only 216/ha in the 1969a plantation. The highest proportion of stumpsinfected pre-harvest occurred in the 1969b plantation (50.5%). The lowest proportion ofstumps infected pre-harvest (38.5%) occurred in the 1984 plantation.2.3.2 RegenerationSpecies Composition Among PlantationsEight conifer species were present in varying proportions in each of the four plantations(Fig. 2). Three species, Douglas-fir, western redcedar and lodgepole pine, constituted over80% of the conifer regeneration.47CONIFER REGENERATION BY SPECIESALL PLANTATIONS COMBINEDIw-Jci-800700600500400300200100I.I.ICw Fd Hw Lw P1 Pw SeSPECIESBIEIGURE 2. Distribution of conifer regeneration by species hi all four plantationsWhile all four plantations had the same eight species present, the relative proportions ofthe total conifer stocking made up by each species varied significantly (Table 14). The mostobvious difference in species composition among the plantations was due to the originalspecies planted on each site. The 1980 plantation was planted to lodgepole pine while theother three were all primarily Douglas-fir plantations with the exception of two areas withinthe 1984 plantation which were also lodgepole pine. A second difference in speciescomposition in the plantations is found between the two age groups (1980/84 versus 1969).The two older plantations have a fairly even distribution of stems among the species while thetrees in the two younger plantations are primarily Douglas-fir or lodgepole pine depending onwhich species was planted.48TABLE 14. Number of conifer trees by species tallied/ha in each of the fourplantationsPLANTATIONSPECIES 1969a 1969b 1980 1984 TOTALBi 45 8 1 1 55Cw 513 937 92 296 1838Fd 751 778 317 1182 302811w 220 209 77 63 569Lw 27 4 8 28 67P1 7 18 1021 329 1375Pw 250 67 82 20 419Se 390 78 11 66 545TOTAL 2203 2098 1610 1985 7897Very few deciduous trees exhibited signs of A. ostoyae infection. Out of a total of1304 paper birch trees examined only 6 or 0.46% exhibited signs of A. ostoyae infection.Similar rates of infection were found for trembling aspen trees with only 3 out of 506 or0.59% infected. In relation to most of the conifer regeneration, these rates were miniscule andthus were not believed to be very helpful in determining infection patterns. However, it isimportant to note that the proportion of western redcedar that was infected was only slightlyhigher than that of the two major deciduous species.Comparison of the Relative Rates of Armillaria ostovae Infection Incidence among ConiferSpecies to a Douglas-fir StandardThere was a significant difference in the proportion of Douglas-fir trees infected amongthe four plantations (Chi-square=94.69, Chi-square critical=7.815). The highest proportionof stems infected occurred in the 1969b plantation (18.0%) followed by the 1980 (12.6%),1984 (7.0%) and 1969a plantation (4.4%). The proportion of the other seven conifer speciesthat were infected were then compared to determine if they varied in accordance with Douglas-fir.The relative rates of A. ostoyae infection incidence were compared among the eightspecies found throughout the four plantations. The proportion of Douglas-fir infected was49used as a standard measure of visible infection incidence for each plantation. The proportionsof trees infected for the other species were compared to the Douglas-fir standard (Table 15).TABLE 15.SPECIES . K-VALUE CHI-SOUARE CRITICAL CHI-SOUAREFd 3028 296 1.0000 - -Cw 1838 12 0.0537 0.450 5.991Pw 419 9 0.2587 4.514 5.991Se 545 11 0.2957 2.900 5.991Hw 569 23 0.3745 8.643 7.815Lw 67 3 0.6146 1.922 3.841P1 1375 167 1.0484 2.589 5.991B! 55 3 1.3840 0.055 3.841Ranking of conifer species compared to a Douglas-fir standard on thebasis of the incidence of Armillaria ostoyae infection in plantationsage 10-25 years#TREE #INFECTPThWestern hemlock was the only conifer species for which the proportion of treesinfected differed significantly from the Douglas-fir standard measure of disease incidenceamong the four plantations. There was no significant difference in the relative rates ofinfection among the four plantations for the other six species compared to the Douglas-firstandard. This lack of significant differences among the plantations meant that a single K-value could be used for each of the six species over the four sites. The K-value quantified therelative frequency of infection for a species. Western redcedar trees, for example, wereinfected only 0.054 times as often as Douglas-fir. Since the Chi-square was significant forwestern hemlock, a single K-value was not used for this species. It appears that the ratio ofinfected western hemlock to infected Douglas-fir differed significantly from site to site. The1969a plantation was primarily responsible this difference. None of the 220 western hemlocktrees tallied in the 1969a plantation were infected, based on above ground evidence. As statedearlier 4.4% of the Douglas-fir trees were infected in this plantation.The differences among species K-values were compared (Table 16). The proportion ofwestern larch trees infected by A. ostoyae was not significantly different from any of the otherspecies. This lack of differences was due primarily to the low number of larch trees in the50four plantations. The proportion of western redcedar trees that were infected was significantlyless than that of all other species except western larch.TABLE 16. Chi-square values from the comparison of the proportions of trees infectedwith Armillaria ostoyae among eight conifer species in four plantations aged10 to 25 yearsFd Cw Hw Lw P1 Pw Se BiFd -Cw 189.482* -Hw 25.873* 34.96* -Lw 3.440 0.0023 2.814 -P1 2.260 38.59* 6.373 1.38 -Pw 20.00* 17.455* 10.188* 2.825 12.71* -Se 18.367* 26.72* 2.416 4.923 7.308 0.042 -Bl 0.566 16.95* 13.568* 2.692 1.607 6.947 8.567*CM-square values followed by * are significantThere was no significant difference in the proportion of trees infected betweenlodgepole pine and Douglas-fir. Western hemlock was included in Table 16 despite thesignificant Chi-square from Table 15. As described earlier, the reason for the significant Chisquare for western hemlock was due primarily to one plantation. In general the proportion ofwestern hemlock that was infected tended to follow the proportion of Douglas-fir infected.Regeneration Species and Disease Status Distribution by Diameter ClassThe species composition and size of each species was naturally dependent on the age ofthe plantation and the species originally planted. Western redcedar and western hemlockregeneration were concentrated in diameter classes less than 7.5cm, with 82.8% and 80.5% ofeach species respectively below that diameter (Table 17). Douglas-fir and lodgepole pine wereconcentrated in larger diameter classes with 53.4% and 74.3% of their stems respectively indiameter classes greater than 7.5cm. These differences in species size were probably due moreto differences in growth rates than differences in age. The western redcedar and western51hemlock regeneration, although smaller than the planted species, would probably have beenestablished shortly after the sites had been burned. Observations in mature stands studied onHunter’s Range (Chapter 3) indicated that suppressed understory western redcedar and westernhemlock trees were approximately the same age as the overstory Douglas-fir.TABLE 17. Conifer regeneration species distribution in percent by diameter class(all plantations)DIAMETER CLASS(cm)SPECIES 1-2.5 2.5-7.5 7.5-12.5 12.5-17.5 17.5-22.5 TOTALB1 10.9 45.5 32.7 7.3 3.6 55Cw 36.3 46.5 14.4 2.0 0.8 1826Fd 7.1 39.5 30.1 14.7 8.6 2950Hw 24.8 55.6 15.9 3.0 0.7 568Lw 14.5 33.9 21.0 21.0 9.7 62P1 0.7 25.0 53.8 19.9 0.7 1375Pw 11.0 41.3 32.1 11.8 3.8 417Se 11.0 55.2 25.9 5.9 2.0 545TOTAL 1141 3180 2276 852 314 7798The greatest proportion (58.9%) of regeneration mortality due to A. ostoyae occurredin the 2.5-7.5cm diameter class (Table 18). The second greatest, 26.5%, occurred in the 7.5-12.5cm diameter class. These two diameter classes together contained 85.4% of allregeneration mortality due to A. ostoyae. These two classes also contained 70.3% of all trees,so it is not surprising that the greatest losses to A. ostoyae are found in these diameter classes.TABLE 18. Armillaria ostoyae severity distribution by diameter class and plantation(% of total in D-class)DIAMETER CLASS(cm)PLANTATION 1.0-2.5 2.5-7.5 7.5-12.5 12.5-17.5 17.5-22.5 22.5-27.5 PLANT TOTAL1984 dying 0.0 0.9 0.4 0.0 0.0 0.0 121984 dead 4.9 7.3 0.7 0.0 0.0 0.0 1021980 dying 0.0 1.4 1.1 1.8 0.0 0.0 201980 dead 8.7 24.4 6.5 0.6 0.0 0.0 1621969b dying 0.2 0.5 2.0 4.7 3.0 2.2 291969b dead 1.5 6.4 11.3 8.6 1.2 0.0 1181969a dying 0.0 0.2 0.3 1.1 0.0 0.0 71969a dead 0.0 1.9 2.2 1.4 0.8 0.0 37TOTAL DYING 1 21 20 20 5 1 68TOTALDEAD 32 247 111 26 3 0 41952The distribution of diseased stems throughout the range of diameter classes wascompared among the four major species (Table 19).TABLE 19. out of thefour majorINFECTED INFECTED DEAD,OTHERDEAD DYING CAUSES TOTAL TREESWestern redcedar0-2.5cm 0.3 0.0 0.0 6632.5-7.5cm 0.5 0.0 0.1 8497.5-12.5cm 0.8 0.8 0.0 26312.5-17.5cm 0.0 2.8 0.0 3617.Scm+ 0.0 3.7 0.0 27Total 8 4 1 1838Dou2las-fir0-2.5cm 10.7 0.0 0.5 2072.5-7.5cm 10.0 1.2 0.2 11657.5-12.5cm 7.6 1.4 0.1 89412.5-17.5cm 6.3 3.5 0.0 43217.5cm 1.2 1.5 0.0 330Total 249 47 4 3028Western hemlock0-2.5cm 3.5 0.7 0.0 1412.5-7.5cm 3.5 0.3 0.6 3167.5-12.5cm 4.4 1.1 0.0 9012.5-17.5cm 0.0 0.0 0.0 1717.Scm+ 0.0 0.0 0.0 5Total 20 3 2 569Lod2epole pine0-2.5cm 10.0 0.0 0.0 102.5-7.5cm 31.2 1.2 0.6 3277.5-12.5cm 6.5 0.4 0.3 72412.5-17.5cm 0.7 1.5 0.0 27217.5cm 0.0 0.0 0.0 9Total 152 11 4 1348Western redcedar had the lowest proportion of stems infected with only 0.65% of thetotal stem count. There was very little difference in the number of western redcedar stemsPercentage of stems infected with Armillaria ostoyaetotal number of stems in each diameter class for theregeneration species53killed among the three smallest diameter classes. Only 8 stems out of 1775 were killed inthose diameter classes (0.45%). Douglas-fir was second only to lodgepole pine in theproportion of stems infected and killed by A. ostoyae. For Douglas-fir, the smallest diameterclass was the most severely affected, with 10.7% mortality. The percent mortality decreasedwith increasing diameter for this species. For western hemlock, the percent mortality wasfairly constant over the three smallest diameter classes. Very few large diameter westernhemlock were tallied. Those that were recorded, were all healthy. Lodgepole pine was thespecies most severely affected by A. ostoyae. In this species, the losses were greatest in the2.5-7.5cm diameter class where 3 1.2% of the trees were killed. The majority of these lossesoccurred in the 1980 plantation, which was planted solely with lodgepole pine. The cause ofthe severe losses in lodgepole pine in this study may not be assigned simply to the species.The results in Table 16 indicated that there was no significant difference in the proportion oftrees infected between lodgepole pine and Douglas-fir. There are confounding reasons for thehigh losses in lodgepole pine. The 1980 plantation had the highest stump inoculum load of thefour plantations, and as pointed out later, had the highest mortality rate from other causes ofthe four plantations.Diameter Class Distribution for each PlantationThe diameter class distribution for each plantation is summarized in Table 20.Naturally, there were no large diameter trees in the two youngest plantations. There was aconsiderable amount of small diameter regeneration in the two oldest plantations. In both ofthe 1969 plantations the 2.5-7.5cm diameter class made up the largest proportion of coniferstocking.54TABLE 20. Percentage of total stems in each diameter class by plantationPLANTATIONDIAMETER CLASS (cm’ 1969a 1969b 1980 19841.0-2.5 8.4 26.1 5.7 16.42.5-7.5 44.7 31.7 26.5 56.37.5-12.5 26.8 19.7 46.1 27.112.5-17.5 12.6 11.3 20.9 0.217.5-22.5 6.0 8.1 0.8 0.122.5-27.5 1.3 2.2 0.0 0.027.5-32.5 0.2 0.9 0.0 0.0TOTAL TREES 2203 2098 1610 1985Comparison of Armillaria ostoyae Severity Among PlantationsThere were significant differences in the incidence of mortality caused by A. ostoyaeamong the four plantations (Table 21).TABLE 21. Percentage of total conifer trees in each plantation killed by or dying fromAnnillaria ostoyae or killed by other causesPLANTATION DEAD DYING OTHER CAUSES TOTAL1969a 1.7 0.3 3.4 22031969b 7.3 1.4 1.1 20981980 10.1 1.2 4.2 16101984 5.1 0.6 0.5 1985TOTAL 453 68 174 7896The 1980 plantation suffered the greatest losses overall with 10.1 % mortality due to A.ostoyae. Mortality from other causes described below was also the highest in this plantationwith 4.2%. The 1969b plantation also had a high rate of mortality with 7.2%. The 1969aplantation had the lowest rate of mortality overall with only 1.7%. The 1969b plantation hadsignificantly more A. ostoyae infection than the unbrushed 1969a plantation. These differencesin mortality rates are perhaps misleading. The 1969a plantation had the lowest rate ofmortality based on the evidence visible. It is quite possible that some trees killed by A.55ostoyae in the first 10 years following establishment of the two oldest plantations would nolonger be visible. Evidence of A. ostoyae caused mortality in the two younger plantations overthe same 10 year period would be more obvious.The 1969b plantation did contain evidence that allowed for an estimate of A. ostoyaecaused mortality at age 15. This plantation was brushed in 1984. At the time of brushing, alldead Douglas-fir stems were cut down (Jim Wright, Salmon Arm Forest District, pers. corn.).The stumps of these Douglas-fir exhibited signs of A. ostoyae infection in the form of mycelialfan impressions. The number of these Douglas-fir stumps provided an estimate of the amountof A. ostoyae caused mortality in this plantation at age 15. Of the 153 (7.3%) trees killed inthe 1969b plantation, 35 (1.7%) were Douglas-fir, killed prior to brushing. It is interesting tonote that the percent mortality in the 1969b plantation prior to brushing was identical to thepresent percent mortality in the 1969a unbrushed plantation.Armillaria ostoyae was not the only cause of death identified in the four plantations.Western white pine was infected to varying degrees by white pine blister rust (Cronartiumribicola J.C. Fisch ex. Rab.) throughout all four plantations. Lodgepole pine was occasionallyinfected with lethal stem galls caused by western gall rust (Endocronartium harknessii [J.P.Moore] Y. Hirat.), Warrens root collar weevil (Hylobius warreni Wood) and a needle blightprobably caused by Lophodermella concolor (Dearn.) Darker. The needle blight was verycommon throughout all of the lodgepole pine, although it was rarely observed as the primarypathogen responsible for tree death. There was a total of 13 dead trees in the four plantationsin the ‘other causes’ category for which the cause of death could not be identified.Spatial Distribution of Infected Regeneration within PlantationsThe distribution of infected trees within each of the plantations was also examined on aper plot basis (Table 22). The 1969a plantation was clearly the least infected in terms of area,with only 34.0% of the plots containing infected trees. In the 1969b plantation, 74.0% of theplots contained trees showing signs of infection by A. ostoyae. The 1980 plantation was most56severely affected by A. ostoyae. In this plantation 85.4% of the plots showed signs ofinfection in the regeneration.TABLE 22. Distribution of plots by increasing frequency of infection for all fourplantationsNUMBER OF PLOTS WITH NUMBER OF INFECTED TREES/PLOT TOTALPLANTATION 0 1-2 3-4 5-6 >6 PLOTS1969a 33 10 6 0 1 501969b 14 10 9 11 6 501980 7 12 11 11 7 481984 13 19 11 3 4 50TOTAL PLOTS 67 51 37 25 18 1982.3.3 The Relationship Between Past Levels of Armillaria ostoyae in Mature Stands andPresent Levels of the Disease in Plantations Established on the Same SitesCorrelation analysis was used to analyze the relationship among 18 combinations ofvariables for all four plantations combined and for each plantation individually (Table 23). Inthese analyses, post-harvest measures of A. ostoyae in stumps included both pre- and post-harvest infections together and so were a measure of the total amount of A. ostoyae on the site.For each individual plantation, the sample size was 50 (48 in the 1980 plantation). The samplesize for the combined plantation analyses was 198. Since the sample size was virtually thesame throughout the four plantations, each of the correlation coefficients in Table 23 can becompared to a single critical correlation coefficient (0.273 for a=O.05, 0.354 for a=0.01).For the analyses involving the four plantations combined, the critical correlation value was0.138 for a=0.05 and 0.181 for a=0.01.57TABLE 23. Correlation coefficients resulting from the correlation of stump andregeneration variables for each plantation individually and for all fourplantations combined compared to critical values (a 0.05 and 0.01)1969a 1969b 1980 1984 ALL SITESTREINFCT vs STMPBEFR 0.117 0.447 0.338 0.088 0.295TREINFCT vs STMPAFTR 0.210 0.365 0.357 0.224 0.282TREINFCT vs ARMSEVB 0.021 0.421 0.049 -0.133 0.225TREINFCTvsARMSEVA 0.140 0.304 0.095 0.017 0.265TREINFCT vs BABEFOR 0.044 0.166 0.163 -0.014 0.157TREINFCT vs BAAFTER 0.044 0.104 0.184 0.148 0.159PROINFCT vs STMPBEFR 0.280 0.53 1 0.362 0.37 1 0.402PRO1NFCT vs STMPAFTR 0.340 0.498 0.385 0.438 0.403PROTNFCTvs ARMSEVB 0.113 0.424 0.028 0.077 0.238PROINFCTvs ARMSEVA 0.200 0.344 0.063 0.138 0.310PROINFCT vs BABEFOR 0.142 0.246 0.184 0.192 0.267PROINFCTvsBAAFTER 0.165 0.214 0.219 0.335 0.294CONSTEMS vs STMPBEFR -0.302 -0.262 -0.269 -0.473 -0.336CONSTEMS vs STMPAFTR -0.372 -0.274 -0.303 -0.357 -0.313CONSTEMS vs ARMSEVB -0.170 -0.126 -0.055 -0.279 -0.187CONSTEMS vs ARMSEVA -0.290 -0.110 -0.095 -0.105 -0.241CONSTEMS vs BABEFOR -0.390 -0.134 -0.158 -0.445 -0.339CONSTEMS vs BAAFTER -0.503 -0.170 -0.192 -0.390 -0.387CRITICAL VALUE a=0.05 ±0.273 ±0.273 ±0.273 ±0.273 ±0.138a=0.01 +0.354 +0.354 +0.354 +0.354 +0.181Comparison of Measurements of Stump InfectionThe numbers of stumps infected both pre- and post-harvest were the stump variables mostclosely correlated with both the number of conifers infected and the proportion of conifersinfected. The correlation r-values were higher for the proportion of conifers infected than forthe actual number of conifers infected for all four plantations individually and combined. Therelationships between infected stumps and infected regeneration were strongest in the 1969bplantation (Fig. 3).58PROPORTION OF CONIFER REGEN INFECTEDvs NUMBER OF INFECTED STUMPS (1969b)0.30.25U * ***NUMBER OF INFECTED STUMPS/PLOTFIGURE 3. The relationship between the number of stumps infected pre-harvest andthe proportion of conifer regeneration infected in the 1969b plantationThe number of healthy conifer stems/ha was most closely correlated with the basal area ofinfected stumps over all four plantations combined. It is not clear why the most highlycorrelated variables for a given plantation are not the most highly correlated over all of theplantations. In the two most severely affected plantations, 1969b and 1980, the number ofinfected stumps was most closely correlated with the number of disease free stems/ha. In theother two less severely affected plantations, the basal area of infected stumps was morestrongly correlated with the number of disease free stems/ha.59The proportion of stumps infected either pre- or post-harvest was not stronglycorrelated with the number of conifer regeneration infected, the proportion of coniferregeneration infected or the number of healthy stems/ha. Both the number of infected stumpsand the basal area of infected stumps were more closely correlated with the regenerationvariables than the proportion of stumps infected.Post-harvest infections were first considered as separate from pre-harvest infections. Thiswas done to determine if there was a difference between the two measures of A. ostoyaeinoculum in their ability to be used as predictors of future regeneration mortality. Thisdivision of inoculum sources was later removed since, logically, those stumps infected preharvest would also have to be considered as being infected post-harvest. The addition of post-harvest infected stumps had very little affect on the correlations (Table 23).Multiple regression, using forward stepwise elimination, confirmed the earlier results inthat the total number of infected stumps per plot was most closely related to the proportion ofregeneration infected with A. ostoyae. All six stump variables were compared simultaneouslyto each of the regeneration variables in each of the plantations. The total number of infectedstumps was the stump variable most closely associated with each of the regeneration variablesin all but the 1969a plantation. In this plantation, the basal area of infected stumps was mostclosely correlated with the number of disease free stems/ha.Effect of Plot Size on the Strength of Relationships Between Infected Stumps and InfectedRegenerationThe effect of plot size on the relationships between the number of infected stumps, bothpre- and post-harvest, and the number of regeneration infected was examined. This was doneby comparing the correlation r-values of STMPBEFR and STMPAFTR regressed onTREINFCT on the full 8m radius plots and those obtained from quarter plot or sectoranalyses. In order to compare the r-values for both plots and sectors, an equal sample size wasrequired. Only the first sector in each plot was used in the analyses. The results of theseanalyses are shown in Table 24.60TABLE 24. Effects of plot size on relationships between the number of infectedstumps and the number of infected regenerationPLANTATION CORRELATION COEFFICIENT n p1969b 0.365 50 0.0121969b only infected* 0.473 36 0.004l9ó9bsectorl 0.022 50 0.8781969b only infected sector l’s -0.230 19 0.3431984 0.224 50 0.0971984 only infected 0.150 37 0.3761984 sector 1 0.333 50 0.0181984 only infected sector l’s 0.277 23 0.201*“only infected” refers to using only plots or sectors with infected regeneration in the analysis.In the 1984 plantation, using sectors rather than full sized plots appeared to improve therelationships between the number of infected stumps and the number of infected regeneration.The r-value for the relationship increased from 0.224 to 0.333 by using sectors. In the 1969brushed plantation it was quite clear that the use of smaller plots did not improve therelationships between infected stumps and infected regeneration. The r-value using full sizedplots was 0.365 while the r-value for sectors was 0.022. As mentioned earlier (Table 23), therelationship between the total number of infected stumps and the number infected trees wasmore closely correlated in the 1969b plantation (r=0.365) than the 1984 plantation (r=0.224).Similar comparisons were made between plots and sectors using only those samplingunits that contained infected trees. The rationale behind this method was that if there was arelationship between proximity of infected trees to infected stumps, it would be clearer if onlythose sampling units containing infected trees were analyzed. Including only the samplingunits containing infected trees in the analysis increased the r-value for the 1969b plantationplot analysis only. The r-value of the sector analysis in the 1969b plantation was decreased byremoving the sectors with no infected trees. In the 1984 plantation, removing sampling unitswith no infected trees reduced the r-values of the relationships for both plots and sectors.612.4 Discussion2.4.1 Comparison of the Relative Rates of Armillaria ostoyae Infection Incidence amongConifer Species to a Douglas-fir StandardThe ranking of conifer species based on incidence of infection (Table 15) differedconsiderably from the ranking that Morrison (1981) suggested. Morrison’s (1981) speciesrankings were based on trees that exhibited signs of being challenged by the disease. Whethera species was ranked as more or less susceptible than another, depended on the reaction of thespecies to the disease. Susceptible species were more likely to succumb to the disease and diewhen challenged than the more tolerant species. The species rankings reported in Table 15were based on the proportion of trees infected of a given species out of the total number of thatspecies. It was not known how many of the apparently disease free trees had been challengedby A. ostoyae. Therefore, it is not valid to compare directly the ranking of speciessusceptibility suggested by Morrison (1981) with that reported in Table 15.Despite the differences in the methods used to rank the species, it appears from thisstudy that the ranking of conifer species susceptibility to A. ostoyae deserves more attention infuture studies. To say that Douglas-fir is more susceptible to A. ostoyae than lodgepole pine(Morrison 1981) is questionable. There was no significant difference in the proportion of treesinfected between lodgepole pine and Douglas-fir (Table 16). There is little reason to believethat one of these species would have been challenged by A. ostoyae more than the other overthe four plantations. Both species would have been exposed to the same inoculum loadsbecause both species were present on all four plantations. A similar argument could be madefor including western redcedar in the same susceptibility class as lodgepole pine. It is possiblethat the plantations were not old enough to express the differences in species susceptibility toA. ostoyae. Morrison et al. (1991a) stated that there was little difference in susceptibility to A.ostoyae among conifer species less than 15 years old. Once again, it is clear that moreresearch is needed into the question of species susceptibility to A. ostoyae.62The Chi-square test results (Table 15) for each of the species, except western hemlock,indicated that the proportion of trees infected for a species did not vary significantly from theproportion of Douglas-fir infected. In other words, in those plantations where the proportionof Douglas-fir infected was high, the proportions of the other species infected were also high.Thus, the differences in site among the four plantations did not affect the relative ranking ofdisease incidence among the conifer species. It was therefore possible to assign a single value(K-value) for each species as a ratio, that described the incidence of disease for that species inrelation to a Douglas-fir standard. It was then possible to test for differences among the K-values. These tests (Table 16) indicated whether or not the proportion of trees infected weresignificantly different between two species. Attaching a statistical significance to the K-valuesallows for a statistically sound, quantitative ranking of conifer species based on the incidenceof A. ostoyae infection.The quantitative ranking of species based on disease incidence has advantages over thequalitative ranking of species susceptibility to A. ostoyae. A quantitative ranking of speciesbased on the incidence of disease could provide forest managers with a tool that they may usein cost-benefit analyses of various management options. If it was known that in a given area,western redcedar was infected 0.0537 times as often as Douglas-fir, that number could beused, along with growth rate information for the two species, in calculations of future yieldsfor a site. It is likely that the K-values reported in Table 15 would change as the stand aged.For example, young western redcedar in the plantations rarely exhibited evidence of infection,while the stump evidence indicated that western redcedar was infected just as often as Douglas-fir and western hemlock. Therefore, a series of K-values over the length of the rotation wouldprobably be required for the calculations of future yields described above. A qualitativeranking such as Morrison’s (1981) aids in decisions involving species choices, but does notprovide the forest manager with values required for more detailed cost-benefit analyses of rootdisease control options.The validity of an overall ranking of tree species susceptibility to A. ostoyae dependson those species having constant relative proportions of trees infected among species63throughout all sites and ages. If the ranking of species suceptibility changes significantly fromsite to site and with age of the stands, it is not possible to determine an overall ranking of anyvalue.2.4.2 The Impacts of Armillaria ostoyae on Species Composition and Successionwithin the ICH zoneArmillaria ostoyae influences the species composition of stands within the InteriorCedar Hemlock zone to a considerable degree. The disease’s influence has been described asincreasing the rate of natural succession in this zone (Morrison 1981). The effects of thisinfluence became quite clear when the differences in species composition between the twoyoungest and the two oldest plantations were examined. The species composition changedover time from being primarily a single species stand at the time of establishment to a muchmore diverse stand by age 25. The two 1969 plantations probably looked quite similar to the1984 plantation about 15 years ago. One interpretation of the data suggests that over time, A.ostoyae has selectively removed the more susceptible Douglas-fir leaving openings for themore tolerant western redcedar, western hemlock and deciduous species. Douglas-fir andlodgepole pine were the two most frequently infected tree species. Western redcedar was theleast frequently infected conifer species in the regeneration and was infected only slightly moreoften than paper birch and trembling aspen.A second interpretation of the data includes the influence of time. The proportion ofstems infected with A. ostoyae was highest in Douglas-fir and lodgepole pine. These specieswould appear to be more susceptible to the disease than western hemlock or western redcedar.The former two species were, however, the species planted and so have been on the sites,exposed to the A. ostoyae inoculum for the longest time. Although the ages of the trees werenot determined, it is likely that the majority of the western redcedar and western hemlock haveestablished in the two oldest plantations sometime after the sites were planted. Differences inspecies composition between the two oldest and the two youngest plantations (Table 14)indicate that there were few western redcedar and western hemlock in the two youngest stands.64Since all of the plantations were on similar sites surrounded by similar mature stands, itfollows that western hemlock and western redcedar would likely ingress into the youngerplantations over time. Thus, the western redcedar and western hemlock on the oldestplantations were probably younger than the Douglas-fir. It is possible, given the same lengthof time on the site as the planted species, that the rates of infection in western hemlock andwestern redcedar could be just as high. The stump data revealed no significant differenceamong these three species in the proportion of stumps infected prior to harvest.The data in Table 19, however, substantiate the claim that Douglas-fir and lodgepolepine are more frequently infected and killed and thus more susceptible than western hemlockand western redcedar. Douglas-fir and lodgepole pine trees in the three smallest diameterclasses, 0-2.5cm, 2.5-7.5 and 7.5-12.5, were significantly more heavily infected than westernredcedar and western hemlock in the same diameter classes. Armillaria ostoyae appears to beaffecting succession in this area of the ICH by selectively removing the early seral species andfavouring the later seral species.The impact of A. ostoyae on stands could also be characterized as slowing or evenreversing succession in stands within the ICH zone. Following major disturbances, such asforest fires, paper birch is often one of the first species to dominate the site. Paper birch alsotends to dominate those site which are most severely infected with A. ostoyae. The briefexamination of deciduous species regeneration, although not nearly as complete as that of theconiferous species in this study, revealed that the proportion of trees infected was significantlyless than that of most conifer species.As mentioned earlier, the majority of stumps within the four plantations examined wereeither western redcedar, Douglas-fir or western hemlock. Deciduous species stumpsdecompose very rapidly in relation to the coniferous species stumps. Thus, very little evidenceof prior deciduous stocking existed in any of the study areas. The loss of deciduous stumpevidence results in an incomplete picture of the past stand’s species composition and possibly aloss of important disease evidence. None the less, it appears that the relative proportions oftree species on the sites were strongly influenced by A. ostoyae. If the species composition of65the original stand was strongly influenced by A. ostoyae, it may be possible to predict the levelof infection in future stands based on the relative proportions of these three conifer speciesalong with the proportion of deciduous trees.The 1980 plantation provides a case in point. The 1980 plantation had the greatest A.ostoyae inoculum load of all of the plantations in terms of infected basal area. The speciescomposition of the original stand on this site was quite different from that of the other threeplantations. There were significantly fewer Douglas-fir stumps and significantly more westernhemlock stumps on this site than on the other three sites. These differences in stump speciescomposition between the four plantations could be due to succession alone. Perhaps theformer stand on the 1980 site was the oldest of the four sites. The Douglas-fir component ofstands in the ICH decreases over time as the species is replaced by the climax species, westernhemlock and western redcedar. Douglas-fir is shade intolerant in the ICH zone thus it doesnot replace itself in an undisturbed stand. Western hemlock and western redcedar are shadetolerant species and are quite capable of growing underneath a Douglas-fir canopy eventuallyreplacing the original stand. Conversely, the low Douglas-fir stump component on this sitecould have been due to lower resistance of this species to the high A. ostoyae inoculum load ascompared to western hemlock and western redcedar.By the time the stands reached maturity (approx 120 years), there was no significantdifference among western redcedar and Douglas-fir in the proportion of trees that wereinfected among all four plantations (Table 7). However, the infected western redcedar stumpswould have probably continued to survive with the disease had the area not been harvestedsince the species has the ability to wall off the basal lesions caused by A. ostoyae with callustissue. The Douglas-fir on the other hand would probably not have fared so well. Douglas-firalso has the ability to grow callus tissue around A. ostoyae infections, but western redcedarappears to be more successful at controlling the disease. As the stump origin data points out,very few of the snag stumps were western redcedar. The majority of them were Douglas-fir.It is likely that western redcedar continues to be more tolerant of A. ostoyae as standsmature than the other conifer species. Western redcedar and A. ostoyae appear to have66reached a relatively stable equilibrium in the ICH zone. The equilibrium between A. ostoyaeand the early seral species such as Douglas-fir and lodgepole pine, is much more unstable.Western redcedar is more tolerant of infections throughout the rotation and is killed much lessfrequently than the early seral species. The differential mortality rates among conifer speciesin response to A. ostoyae infections, allow western redcedar to maintain its position in standswhile other species are killed. The equilibrium between A. ostoyae and the early seral speciestends to shift in favor of the disease with ease. The disease may be triggered by factors suchas logging, insect attack or other diseases. The insect and disease factors in the ICH tend toaffect the early seral species more than western redcedar, putting them at a furtherdisadvantage in this zone. Early seral species in the ICH zone may be killed by sprucebudworm (Choristoneura occidentalis Freeman), Douglas-fir bark beetle (Dendroctonuspseudotsugae), spruce beetle (Dendroctonus rufipenis Kirby), mountain pine beetle(Dendroctonus ponderosae Hopk.), and white pine blister rust (Cronartium ribicola Fisch.)among others. Any one of these insects or diseases could possibly be responsible for shiftingthe equilibrium in favor of the disease triggering A. ostoyae to become aggressive in standsalready infected with the pathogen (Kuihavy et al. 1984, Cobb 1989, Filip 1989b). Westernredcedar was rarely killed by any disease or insect other than A. ostoyae, based onobservations from the Hunter’s Range study area (Chapter 3). The effects of A. ostoyae instands within the ICH favor the survival of western redcedar over other conifer species and,therefore, influence the path of succession in this zone.2.4.3 The Relationship Between Past and Present Levels of Armillaria ostoyaeon a SiteThere are inherent difficulties associated with analyzing data sets that contain a largenumber of zeros, particularly when the zeros are associated with the dependent variable. Alarge number of zero values for the dependent variable creates a situation where part of thedata belong to only two classes. In such cases, the dependent variable is referred to as beingdichotomous. Without a normal distribution throughout, the data fails to meet the basic67requirements of simple linear regression. Many of the trees in the four plantations werehealthy and occasionally entire plots exhibited no visible signs of A. ostoyae infection.Therefore, these plots would have a proportion of visibly infected regeneration of zero. At thesame time, these plots often contained stumps which showed signs of infection. Very fewplots were completely disease free in terms of infected stumps. The result is that therelationship between the various measures of stump infection and the proportion ofregeneration infected was weak (Fig. 3). The weakness of these relationships was due in largepart to the number of zeros in the infected regeneration data. With the zero values removed,the relationship between the proportion of regeneration infected and either the basal areainfected or the number of stumps infected was stronger.Aldrich and Nelson (1985) reviewed the use of LOGIT and PROBIT models fordealing with dichotomous dependent variables. The use of such models for improving thestrength of the relationships between A. ostoyae infections in stumps and regeneration wasconsidered. The result of using these models would be, at best, a more accurate prediction ofthe amount of regeneration mortality in a plantation given that some mortality had alreadyoccurred. These models would not increase the probability of accurately predicting the amountof regeneration mortality based on the inoculum load in the previous stand alone. The value ofsuch models was not considered to be great in this case. One of the primary objectives of thisstudy was to develop methods for predicting the degree of loss to A. ostoyae in a plantationbased on the amount of infection in the mature stand prior to harvest. A prediction ofexpected future regeneration mortality at the time of harvest would aid forest managers indeciding what to prescribe for the site in order to improve productivity of the future stand. Ifan accurate prediction of regeneration mortality cannot be made until there is already somemortality occurring in the plantation there is little point in making such a prediction. It is toolate at that stage in plantation development to do much to improve the situation. Steps must betaken much earlier, at the pre-harvest stage, in order to deal effectively with the disease.It is important to keep in mind that the conclusions drawn from a plot analysis may notbe directly applied on a plantation basis. If, for example, a plot contained no infected stumps,68it is still quite possible that trees growing on that plot may become infected. The areasurrounding the “clean” plot could easily contain infected stumps. However, if an entireplantation was free of infected stumps, the probability of a tree becoming infected on thatplantation would be almost nil. In order to predict the average mortality for a given plantationbased on pre-harvest infection levels, a large number of plantations would have to be sampled.Such a study would require a pre-harvest survey of root disease in each plantation, withfollow-up surveys of regeneration survival and root disease activity. The results of such astudy would provide forest managers with the information they need to plan more effectiveroot disease control programs.The relationship between the basal area and number of infected stumps, and theproportion of regeneration infected was not strong on an individual plot basis. Therelationship was similarly not clear on a plantation wide basis. Regeneration mortalityattributable to A. ostoyae was greatest in the 1980 plantation where 10.1 % of the stems werekilled by the disease. The disease was quite evenly distributed over the plantation. Only 14%of the plots in the 1980 plantation exhibited no evidence of infection in the regeneration. Theremaining 86% of plots contained at least one infected tree each. The highest number ofinfected trees in a plot in the 1980 plantation was 15. The majority (83.1%) of the stumps inthe 1980 plantation were infected with A. ostoyae either before or after harvest. The speciescomposition of the regeneration may have contributed to the high levels of mortality in the1980 plantation. However, the results reported in Table 16 indicated that there were nosignificant differences between lodgepole pine and Douglas-fir in the relative proportion oftrees infected. Regardless of possible confounding species effects, this plantation had thehighest cross sectional area of stumps infected with A. ostoyae prior to harvest and suffered thehighest losses following plantation establishment.The relationship was not so clear in the other plantations. The two 1969 plantationshad different treatment histories and therefore the proportions of regeneration infected couldnot be legitimately compared between the two. The differences in the number of infectedstumps between the two were not great. Furthermore, any effects on regeneration survival due69to differences in initial inoculum loads would have been confounded by the effects of thebrushing treatment in the 1969b plantation. The effects of the brushing treatment in the 1969bplantation are discussed later. The 1984 plantation had the highest number of infected stumpsof the four sites. This plantation also had one of the highest proportions of stumps infectedwith A. ostoyae (70.6%). However, the proportion of regeneration infected on this plantationwas only 5.7%, second only to the 1969a plantation for low incidence of infection. Perhapsthe inconsistency of the relationship between infected stumps and infected regenerationdemonstrated by this plantation was due primarily to the age of the plantation. It is possiblethat with another 5 years the relationship between infected stumps and infected regenerationwould be clearer, as in the 1980 plantation. Whether or not more time would accentuate therelationship is not known. The results from this plantation illustrate the high variabilityassociated with the relationship between infected stumps and infected regeneration. The abilityto predict future plantation losses based on the amount of infection in mature stands remainsquestionable.2.4.4 The Proximity of Infected Regeneration to Infected StumpsIt is apparent that the proximity of regeneration to infected stumps does not increasetheir likelihood of being infected by age 25. Conceivably, if the spatial distribution of infectedstumps did have a large influence on the distribution of infected regeneration, examining theplots on a more detailed scale should have improved the correlations between the twomeasures. Smaller plots would tie together more closely the infected stumps with infectedregeneration, if in fact they were closely associated with each other. The correlations betweenthe number of infected stumps/sector and the number of infected regeneration/sector wouldthen be expected to increase. The strength of the correlations did not improve over thosefound using the full sized plots in the 1969b plantation (Table 24). This suggested that thetree’s proximity to infected stumps no longer had a large influence in determining theprobability of that tree becoming infected. Morrison (Forestry Canada, unpublished data)70stated that new infections originated from contacts with diseased trees rather than infectedstumps in stands older than 20 years. The results from this study support that claim.It appeared that stumps were still functioning as the primary source of inoculumresponsible for infecting the regeneration at age 10. Examining the 1984 plantation on a sectorby sector basis improved the correlations between the number of infected stumps and thenumber of infected regeneration. The correlation r-values were higher in the sector analysisthan in the plot analysis in this plantation. These results suggest that reducing the maximumdistance of a tree from an infected stump increased the probability that that tree would beinfected.Why the sector relationship was more significant in the 1984 plantation than the 1969bplantation is, at first, not clear. The differences in age would be the most obvious reason. Upto age 10, infected stumps were still acting as the primary inoculum source, thus infected treesin the 1984 plantation were still associated with the infected stumps. The sector analysissuggested that in the 1969b plantation the stumps were no longer the primary inoculumsources. If infected trees in the 1969b plantation were at one time associated with infectedstumps, this no longer seemed to be the case. It would appear that the evidence of treemortality that was closely associated with the infected stumps in the past was no longer visiblein this plantation. The fact that the 1969b plantation was brushed in 1984 provides someimportant information regarding the question of proximity to infected stumps. Douglas-firtrees that had been killed by disease up to 1984 were cut down during the brushing treatment.The evidence of these dead trees was still present on the 25 year old 1969b plantation. Sincethe evidence of past mortality for the 1969b plantation up to age 15 was still quite visible, it islikely that much of the evidence of past A. ostoyae caused mortality was still present on the1969b plantation. Why then was there not enough evidence of infected trees surrounding theinfected stumps in the 1969b plantation to make the sector analysis more significant on thissite? The answer most likely lies with the brushing treatment. Following brushing, A. ostoyaewas re-activated and spread rapidly through the freshly cut stumps of the western hemlock,western redcedar and birch regeneration. Any of the Douglas-fir crop tree root systems that71were in close association with the infected brushed stems were then exposed to a much greaterinoculum potential. An increasing proportion of the Douglas-fir trees have died sincebrushing. These killed trees were not infected through contact with A. ostoyae colonizedstumps from the original stand, but most likely by contact with colonized-brushed stumps. Itis not surprising, therefore, that the relationship between infected stumps and infectedregeneration in this plantation was not improved by sector analysis.These results have important implications for guidelines concerning silviculturalsurveys in the ICH zone. The current guidelines state that any tree within 3 meters of aninfected stump can not be considered as free to grow. The results from this study indicatethat, in 25 year old plantations, trees within 4m from an infected stump are no more likely tobe infected than those trees within 8m if the site has been brushed. In 10 year old plantationsthere was a stronger relationship between infected trees and infected stumps. Thus, thevalidity of the 3m rule depends on the age of the plantation and the treatments that have beenconducted on the site. A blanket treatment over all plantations within the ICH zone is notappropriate.2.4.5 The Impacts of Brushing Stands Infected with Annillaria ostoyaeThe negative impacts of brushing and spacing on the survival of crop trees in areasinfected with A. ostoyae were quite clear after comparing the two oldest plantations. Thesetwo plantations were similar in many respects. They both had similar initial inoculum loads asmeasured by the basal area of stumps colonized by the fungus, both prior to and followinglogging. The unbrushed plantation had 44.36m2/haof infected stump basal area compared to37.95m2/ha in the brushed plantation. The species composition of the regeneration was verysimilar between the two. The location, slope and aspect of both plantations was also verysimilar. There were two major differences between the plantations that may have influencedthe development of A. ostoyae development on these two sites. The first difference was thedate of harvest. Although both plantations were established in 1969, the original harvest dates72were not the same. The 1969 brushed plantation was logged in 1965 while the unbrushedplantation was logged in 1968. The delay in planting the 1969b plantation probably createdthe need to brush that stand. The delay in replanting would also be expected to decrease theinoculum potential of A. ostoyae on the site. The planting lag would leave the disease withfew suitable new hosts for the 5 year period between logging and planting. The seconddifference, obviously, was the fact that one of the plantations was brushed while the other wasnot. A summary of the similarities and differences between these two plantations is found inTable 25.TABLE 25. A comparison of the two 1969 plantations1969a 1969bYEAR LOGGED 1968 1965YEAR PLANTED 1969 1969YEAR BRUSHED - 1986BASAL AREA INFECTED 44.4m2/ha 38.0m2/haDEAD & INFECTED STEMS/HA> 12.5cm DIAMETER 8/443 (1.8%) 40/472 (8.5%)% OF PLOTS INFECTED 34 72The greatest losses have obviously occurred in the brushed plantation. The Douglas-firtrees that had been killed up to the time of brushing were cut down during the brushingtreatment (Jim Wright, Salmon Arm Forest District, pers. corn.). The stumps from these treeswere still visible on the site. Based on this stump evidence, 35 Douglas-fir stems/ha had beenkilled by A. ostoyae up to 1984. Two of these dead Douglas-fir trees were over 12.5cm indiameter. Twenty-one stems/ha of Douglas-fir over 12.5cm diameter have been killed by A.ostoyae since the time of brushing in the 1969b plantation. At the end of the 1993 growingseason, 17 stems/ha of Douglas-fir over 12.5cm diameter were infected and dying. Judging bythe reduced intemodal length of some of the trees, it was likely that these infected trees hadbeen suffering from the effects of A. ostoyae infection for several years. Other infected treesshowed no signs of reduced growth in the past year. Overall, it appeared that the majority ofinfected and dying trees had probably been influenced by A. ostoyae over approximately the73past 4 years. Thus the rate of mortality in the brushed stand appears to be increasing. In 9years, 21 stems/ha over 12.5cm DRC were killed, while in the past 4 years, 17 stems/ha over12.5cm DRC were infected and dying. Conversely, the rate of mortality in the unbrushedplantation was much lower. Only 5 stems/ha over 12.5cm DRC had been killed. The timeperiod over which these trees were killed was difficult to estimate, but it was probably quitesimilar to the 1969b plantation (approx. 10 years). Only 3 stems/ha over 12.5cm DRC wereinfected and dying over the same 4 year period described for the 1969b plantation.There were also large differences between the two plantations in the number of plotsthat contained infected trees. As shown in Table 23, 34% of the plots in the 1969a plantationcontained infected trees. In the 1969b plantation 72% of the plots contained infectedregeneration. In this same plantation, 34% of the plots contained 5 or more infected treescompared to only 2% in the unbrushed plantation.It appears that the disease-host equilibrium discussed in the previous section has beenreached in the unbrushed plantation. Brushing the 1969b plantation has seriously disrupted thepath to equilibrium in this plantation. It appears possible that the disease-host equilibrium maynot be reached in time to salvage a productive stand by the end of the rotation period. AsMorrison (1981) suggested, brushing and spacing should not be carried out in Douglas-firstands infected with A. ostoyae. The results of this study strongly supports that suggestion.The low rate of mortality in the unbrushed, 1969a plantation suggests that perhaps thedisease is reaching an equilibrium with the host by age 25. It is also possible, that thisplantation has always had a low rate of infection in the regeneration and the rate is notslowing. Ideally, a comparison between the 1969a plantation and the 1984 plantation couldprovide more definitive answers as to whether or not the impacts of the disease decline by age25. However, it is more difficult to compare the 1984 plantation to the 1969a plantation thanit was to compare the two 1969 plantations. The ten-year-old 1984 plantation had many sitefactors in common with the 1969a plantation, but there were some important differences. Forexample, the 1984 plantation had considerably more infected stumps/ha than the 1969a74plantation. These differences in initial inoculum load made it difficult to compare rates ofinfection in the regeneration.Even if the rate of infection has always been low in the 1969a plantation, thisinformation in itself is reason for some optimism. The 1969a plantation contained 216infected stumps/ha which represented an infected stump basal area of 44.4m2/ha, yet theconifer stocking on that site now is generally quite healthy. There is some mortality stilloccurring due to A. ostoyae and it will most likely continue. There are still 435 healthyDouglas-fir stems/ha over 12.5cm diameter and considerable western hemlock and westernredcedar ingress to fill in the gaps. The probability of this stand reaching maturity appearsquite high despite the former levels of A. ostoyae infection on the site. The prognosis for thestand that was re-entered is obviously not so optimistic.75CHAPTER 3 ARMILLA1?JA OSTOYAE DISTRIBUTION AND SEVERITY IN THE ICHAND 11W BIOGEOCLIMATIC ZONES3.1 IntroductionArmillaria ostoyae causes significant damage to forests throughout the southern interiorof B.C. The ICH biogeoclimatic zone is one of the most severely affected ecological zones inthe province. The IDF zone is also affected, but generally to a lesser degree. The areasinfected with A. ostoyae in these two zones are generally believed to be quite large.McDonald et a!. (1987) suggested that disease incidence is greatest in host populationsin the transition zones between climax species. They hypothesized that the species in thesezones are less adapted to the sites and therefore less resistant to disease. If this is the case,intraspecific variations in host resistance would be most visible in these transition zones. Thedifferences in site conditions would conceivably place extra stress on those species leastadapted to the specific conditions on that site. In their review of Armillaria spp. research, Kileet a!. (1991) stated that intraspecific variation in host resistance to A. ostoyae has not beeninvestigated. Interspecific variation in host resistance has similarly received little attention.Morrison (1981) reported a qualitative ranking of tree species susceptibility to A. ostoyae forthe southern interior of B.C. As discussed in the previous chapter, more research is requiredconcerning the ranking of susceptibility to A. ostoyae among the species present in the ICHand IDF zones.The interaction between A. ostoyae and other abiotic and biotic factors within thesouthern interior of B.C. has not been studied thoroughly. Several of the following factorscould have an influence on the expression of this disease: the biogeoclimatic siteclassification, site index, elevation and disturbance history of the sites. Armillaria ostoyaedistribution has been associated with particular forest habitat types using the Daubenmire forestvegetation classification system in the Northwest United States (Byler et a!. 1987 and 1990,McDonald 1990, Williams and Marsden 1982). These studies emphasize the dynamic nature76of the interaction between Arinillaria species and natural forest ecosystems (Kile et al. 1991).To date, there has been little work undertaken to link the incidence and severity of A. ostoyaeto the biogeoclimatic ecosystem classification (BEC) system used in B.C. The BEC system,developed by Dr. V.J. Krajina and his students, incorporates primarily soil, climate, andvegetation data (Meidinger and Pojar 1991). This system of ecological classification is usedby the majority of forest managers in B.C. If the probability of a stand being infected with A.ostoyae could be linked to this classification system, it would aid forest managers by providinga simple planning tool for dealing with this forest pathogen.The BEC system uses a fairly wide range of factors to describe the conditions on a site.Other factors, such as elevation and site index, may also influence the behavior of A. ostoyaeon sites within the same BEC site identification. Hobbs and Partridge (1979) found norelationship between elevation and the incidence and severity of A. ostoyae in northern Idaho.Williams and Marsden (1982) found the opposite in the same area of northern Idaho.Forest management practices have a large influence on the expression of A. ostoyae.Partial cutting practices, in particular, have been blamed for increases in A. ostoyae activity innumerous studies carried out in the Northwest United States (Filip and Goheen 1982, Byler etal. 1987, Shaw et al. 1976). Much of the available information is observational, however,and little experimental work has been done on the effect of management practices on diseaselevel (Kile et a!. 1991). The effects of selective harvests in A. ostoyae infested areas in thesouthern interior of B.C. has received very little attention. Much of the southern interior hasbeen selectively harvested over the past century either for poles or railroad ties. The impact ofthis disturbance on the behavior of A. ostoyae in these stands has not been studied.The relationship between A. ostoyae and P. weirii, the two most serious root diseasepathogens in the southern interior of B.C., is also not well understood. Armillaria ostoyae isbelieved to cause damage in conjunction with other root rot pathogens of mixed coniferousforests, particularly with P. weirli (Filip and Goheen 1984, James et a!. 1984, Williams andLeaphart 1978). The existence of the relationship between these two diseases is not acceptedby all workers in this field of forest pathology. Hansen and Goheen (1989) attributed the77associations between these two diseases to chance and to primary secondary relationships.Whether or not there is a relationship between these two disease in the southern interior is notknown.The abiotic and biotic factors described above may or may not influence the activity ofA. ostoyae in the southern interior. If there is an impact on A. ostoyae from any one of thefactors, knowledge of the corresponding impact of A. ostoyae activity on timber volume wouldbe very beneficial. With the exception of Bloomberg and Morrison’s 1989 study, the impactsin terms of forest productivity loss have likewise not been well addressed in the southerninterior. No empirical studies to date have attempted to correlate A. ostoyae severity in amature stand with the corresponding loss in coniferous timber volume.One of the objectives of this study was to gain a better understanding of the ecology ofA. ostoyae in the ICH and IDF biogeoclimatic zones. This study provides an estimate of thedistribution of A. ostoyae within both zones and in the transition between the two. Differencesin site characteristics within the study locations provided an opportunity to examineintraspecific variation among several conifer species to A. ostoyae. The variety of speciespresent in the study area allowed for an examination of interspecific differences in incidence ofA. ostoyae infection among eight conifer species. This research also examined the possibilityof using the BEC system for predicting the probability of stands in the IDF and ICH zonebeing infected with A. ostoyae. The relationship between A. ostoyae severity and site wasviewed at two levels of classification within the BEC system.In addition to studying the impacts of site, as defined by the BEC system, thepossibility of elevation and site index affecting root disease in the ICH and IDF zones ofsouthern B.C. was also examined.Another objective of this research was to examine the relationship between priordisturbance and A. ostoyae activity in the ICH and IDF biogeoclimatic zones.After the effects of the various factors on A. ostoyae severity were examined, therelationship between A. ostoyae severity and both paper birch and conifer volume wereanalyzed.783.2 Methods3.2.1 Study AreaTwo separate areas were sampled in this study. The first was located on the westslopes of Hunter’s Range, directly to the east of Mara Lake, close to Sicamous, B.C. Thisarea straddles the transition between the Interior Cedar Hemlock (ICH), and the InteriorDouglas-fir (IDF) biogeoclimatic zones. Within this study area, the predominant variants ofeach of these zones were the Shuswap moist warm ICH (ICHmw2) and the Shuswap moistwarm IDF variant (IDFmwl) respectively (Lloyd et al. 1990). The Hunter’s Range siteconsisted of approximately 4000ha of “relatively” undisturbed forests (i.e, few recentcutbiocks). A second site was chosen on the northwest slopes of the Larch Hills betweenSicamous and Canoe, B.C. This study area was similar in many respects to the Hunter’sRange area including the aspect, elevation, disturbance history, and biogeoclimatic site types.The two study areas are less than 12km apart.The majority of the area on both study sites was burned in the late 1800’s as a result ofrailroad construction. Thus, the age of most of the stands was approximately 120 years.Stands throughout the Shuswap region have also been selectively logged for poles or railroadties since the early 1900’s. Evidence of past logging disturbance was present throughout muchof the two study areas. More precise dates for the history of these disturbances were notavailable. The predominant forest cover type was Douglas-fir at or near 120 years of age.3.2.2 Sampling DesignThe sampling design for the Hunter’s Range site consisted of 202 - 8m radius plots(0.O2ha) laid out in strip pairs forming a grid. The strip pairs were laid out 400m apart alongthe contour. The plots were located at 200m intervals along the strips which ran east-westagainst the contour. This orientation captured the greatest elevational variation to test ifelevation had any influence on disease expression. Plots were located by tight chaining from79known tie points found on 1:15000 scale Forest Cover Maps (82L075, 82L076, 82L065 and82L066) produced by the Inventory Branch of the Ministry of Forests. Since the siteidentifications for the plots could not be made until they were located, there was an unevennumber of plots in each of the biogeoclimatic zones and among the site types.Thirty-two 8m-radius plot were located using similar methods to those used in theHunter’s Range study for the Larch Hills site.3.2.3 Sampling ProcedureOnce the plot center was located, a 30m Eslon tape was used to establish the perimeterof the 8m radius plots. The perimeter of the plots were flagged with florescent ribbon.Within each plot, the species, diameter at breast height, height and disease code for each livingtree and dead standing tree were recorded.The disease code was determined by examining each tree for the presence of diseasesigns and symptoms. Evidence of A. ostoyae infections included basal resinosis, mycelial fansbeneath the bark, healed over basal lesions and thinning crowns. Any trees exhibiting crownsymptoms were closely examined to determine if A. ostoyae was the causal agent. Rootexcavations were conducted only on those trees exhibiting crown symptoms. Trees with basalresinosis visible above the ground at the base or on exposed roots were examined for thepresence of mycelial fans beneath the bark. These examinations were conducted with an axeand involved chopping into and prying off the resin soaked bark. Healed over basal lesionswere identified by their appearance and by the hollow sound they made when hit. Healed overbasal lesions typically form flattened portions on the bole directly above an infected root.Each tree was also tapped around the circumference of the root collar with the back of the axe.Those trees with flattened portions on the bole, and those that sounded hollow when hit wereexamined further by removing the bark at the area in question. If there was no above groundevidence of A. ostoyae infection present on a tree, the tree was classified as clean. Trees wereassigned a disease code as follows:800 - no sign of infection2 - healed over basal A. ostoyae lesion3 - active mycelium, or recent A. ostoyae mortality.Based on the individual tree examinations within each plot, a preliminary diseaseseverity rating was recorded for each plot. The plot disease severity rating was based on thefollowing scheme:o - no evidence of disease in plot1- evidence of disease on trees < 10cm dbh or dead and down trees only2 - presence of disease code 2 trees > 10cm dbh within plot3 - presence of disease code 3 trees > 10cm dbh within plot.Trees exhibiting crown symptoms and trees within disease centers were also examinedfor the presence of P. weirii. Evidence of P. weirli consisted of ectotrophic mycelium onroots, reddish-brown staining of the heartwood, and thinning rounded crowns. Data on thisroot pathogen were collected in order to describe relationships between this disease and A.ostoyae.A representative tree height was measured for each species present in each plot. Thisheight was used to visually estimate the remaining tree heights within the plot. The height anddiameter data were then used to determine the volume for each tree using whole stem cubicmeter volume equations for each species (British Columbia Forest Service Inventory Division1976).The plot site index was determined by first measuring the height and age of a dominantDouglas-fir within each plot. This information was then used in conjunction with Thrower andGoudie’s (1992) site index tables for interior Douglas-fir to find the site index value for eachplot. If no suitable Douglas-fir was present within the plot, a sample was chosen from nearby.Evidence of prior logging disturbance was recorded using the following classification:0 - no disturbance1- disturbance present outside plot radius (i.e, stumps or skid trails)2 - stumps present within plot.81The species and diameter at root collar was recorded for any stumps within plots.Each plot was classified by biogeoclimatic site types using the field guide for theKamloops Region (Lloyd et al. 1990). Site type classifications were based on the major treespecies forming the forest canopy, as well as plant indicator species present in the understory.The elevation of each plot was determined by transcribing the plot locations from the1:15000 scale Forest Cover Map to 1:20000 scale contour maps provided by the Ministry ofForests District Office in Salmon Arm B.C. These maps allowed for elevation estimations tothe nearest 5m.The last step in data collection involved determining a final A. ostoyae severity ratingfor each plot. The basic logic of the disease severity rating was as follows:if no evidence of A. ostoyae was present within the plot,then A. ostoyae rating = 0;if there was evidence of A. ostoyae on standing trees withinthe plot, then determine the proportion of class 2 and class 3 infected trees out of totalnumber of conifer trees and assign severity rating based on the following ratings:Proportion of total conifers/plotwith class 2 or 3 infections Class 2 Class 30.00-0.10 +1 +30.11-0.20 +2 +60.21 - 0.30 +3 +90.31 - 0.40 +4 +120.41- 0.50 +5 +150.51 - 0.60 +6 +180.61 - 0.70 +7 +210.71 - 0.80 +8 +240.81 - 0.90 +9 +270.91 - 1.00 +10 +30Class 3 infections represented more active and aggressive A. ostoyae activity and thuswere given a higher severity rating than those trees with class 2 infections. Class 2 infections,as mentioned earlier, were those infections that had been actively resisted by the host tree withno signs of active mycelium. The plot severity rating was the sum of the values for class 2and 3 infections.82This root disease severity rating formed the basis for comparisons among the plots fromdiffering site types, elevations, disturbance histories, site indices and volume classes. The plotseverity ratings formed the dependent variable in this study. A similar disease rating wasdetermined for P. weirli.Any root disease survey involves a compromise between the accuracy of the estimatednumber of infections and the time available for surveying. The most accurate estimates of thenumber of A. ostoyae infected trees in a stand would require complete excavation of each tree.Such estimates would be impossible on anything but the smallest scale due to the effort andcost involved in excavating the trees. In order to accurately estimate the extent of A. ostoyaeover a large area, more time must be allotted to covering the ground rather than digging it up.The objectives of this study dealt primarily with the relationships between site and A. ostoyae;thus, the emphasis was placed on covering a large area with a large number of plots.A follow-up study was conducted on the Larch Hills in order to determine moreaccurately the number of infected trees in mature stands within the ICHmw2. A subsample of10 plots was taken within an area with similar site characteristics to that of the other studyareas mentioned earlier in Chapter 3. The results from the subsample demonstrate moreclearly the relationship between the amount of A. ostoyae found and the effort required to findit. The subsample consisted of plots located in the mature stands surrounding the plantationssampled in Chapter 2. Data from the same 10 plots were used in Chapter 2 to ensure that themethods used in the plantation study were not over-estimating the number of stumps infectedpre-harvest. In Chapter 3 these data are used to demonstrate how more time and effort spentin excavating roots can reveal more A. ostoyae infections.Plot locations were established using the same methods as those described earlier (3.2.2Sampling Design). Within the plot perimeter, the diameter at breast height and species ofevery tree was recorded. The roots of each tree were then excavated to a distance of 30cmfrom the root collar and a depth of approximately 10cm into the mineral soil. The bark of theexposed roots was then removed using an axe to determine whether or not mycelial fans of A.ostoyae were present. It was decided at the time of data collection whether or not a given83infection would have been tallied using the methods from the Hunter’s Range study. Thisdecision was based on the amount of excavation required to find the infection. Those treeswith above ground symptoms would have been recorded in the 1992 field season. Those treeswith infections below ground but within 30cm of the root collar would have only beenrecorded in the 1993 field season. The result was a comparison of the ability of the twomethods to detect A. ostoyae infections in a stand.3.2.4 Data AnalysisThe data was first entered in Quatro-Pro (Version 3.01. 1991) and later transferred toSYSTAT (Version 5.03. 1991) for statistical analysis. The data analysis was divided into twosections; the first section dealt with individual trees and the second with plots. Thesignificance level was set at a =0.05 for all tests in Chapter 3.Individual treesDifferences among tree species in terms of relative proportions infected with A. ostoyaewere examined first using the Hunter’s Range data. The Chi-squared statistic was used withcontingency tables, to test the significance of differences in tree species distributions betweenzones and the frequency of infections among species within zones. For both of thesecomparisons, a Chi-square test was first performed on all conifer species combined. If the testindicated significant differences among the species, further tests were performed for pairwisecomparisons between the species. Those species with similar proportions were tested using 2by 2 contingency tables to determine if two species had significantly different proportions ofinfected trees. This method of comparing species was continued until all possible speciescomparisons had been tested. Differences within species in the proportion of trees infectedbetween zones were also tested using 2 by 2 contingency tables (Chi-square critical = 3.841,84one degree of freedom). A similar procedure was used to determine if the frequency ofinfections within a species was significantly different among site units within a zone.The same methods were then used on the Larch Hills data (32 plots) to determine if thedifferences among species in the proportion of trees infected, were consistent between the twoareas. In particular these data were collected to further examine the incidence of infection forwestern larch compared to the other conifer species in the IDF and ICH zones.The data from both the Hunter’s Range site and the Larch Hills sites were thencombined to examine the differences in disease incidence within species, between the ICH andIDF zones in both sites. The proportion of trees infected for five major species werecompared between the Hunter’s Range and Larch Hills sites. The comparisons were testedwith 2 by 2 contingency tables and the Chi-square statistic.The species composition was also compared between the Larch Hills and Hunter’sRange. The proportion of the total trees comprised of each of the five major species werecompared with a Chi-square test. This comparison was made in order to determine if thedifferences in A. ostoyae infection incidence between the two study areas were due todifferences in species composition.The incidence of infection among three of the species (western redcedar, western larchand lodgepole pine) were compared to the incidence of infection for Douglas-fir. The methodsused for this analysis were the same as those used to test for differences among the coniferregeneration species in Chapter 2 (pages 30-31). Douglas-fir was chosen as a standardmeasure of disease incidence for four areas; Hunter’s Range ICH, Hunter’s Range IDF, LarchHills ICH, Larch Hills IDF. These methods were designed to determine a quantitative rankingof conifer species based on incidence of A. ostoyae infection in mature stands. There were notenough of the other species (western hemlock, Engelmann spruce, western white pine and subalpine fir) to be included in this analysis.The incidence of A. ostoyae infections in conifer species were also compared among thebiogeoclimatic site units within each of the zones. The proportions of infected trees out of thetotal for each species were compared among site units using a Chi-square test. If the85proportion of trees infected differed significantly among site units, further pairwise comparisonChi-square tests were used to determine which site units had significantly differentproportions.PlotsPlot data were summarized by biogeoclimatic zone and site unit, site index, elevation,disturbance history and disease severity. Plot data combined the information for all speciespresent in the plot into single values for conifer volume, conifer basal area, deciduous volumeand basal area and A. ostoyae severity. Tree volumes were calculated for each plot (m3/ha) byspecies using whole stem cubic meter volume equations (Ministry of Forests, InventoryBranch, 1976). Volumes for each plot were separated into the following categories: conifervolume, birch volume and total volume in cubic meters. Basal area (m2/ha) was alsosummarized for each species.The relationships between A. ostoyae severity and disturbance, elevation, site index,and Phellinus severity were compared using correlation analysis to determine which factorswere most closely related to the disease. The influence of each of these factors were alsoexamined individually.Armillaria ostovae severity and disturbanceThe relationship between logging disturbance levels and A. ostoyae severity was testedwith a one-way analysis of variance (ANOVA). Disturbance histories were divided into threecategories (page 80) to determine if there were differences in the mean plot A. ostoyae severitydepending on the degree of disturbance.86Root disease severity and elevationThe relationship between disease severity and elevation was tested for both A. ostoyaeand P. weirli. Correlation analyses were used to determine if there was a significantrelationship between the two root diseases and elevation.Armillaria ostovae severity vs Biogeoclimatic site unitThe relationship between A. ostoyae severity and biogeoclimatic site unit was testedwith a one-way analysis of variance (ANOVA). This test determined if there were significantdifferences among site units in the mean plot severity rating within each zone. The same testwas used to determine if there were significant differences between the two zones.Armillaria ostovae severity vs timber volumeThe impact of A. ostoyae severity on plot volumes was examined using three differentvolume measures (conifer volume, birch volume, and total volume). These relationships wereanalyzed using a one-way ANOVA in each of the zones individually and both zones combined.Armillaria ostoyae severity ratings were grouped into 4 classes. The first class, Class 0included only disease free plots. Severity ratings from 2.0-7.99 were in Class 5, ratings from8.00-13.99 were in Class 10, and ratings from 14.00-21.00 were in Class 15. The meanvolumes were then compared between the 4 classes in both the ICR and IDF zones. Tukeytests were used to identify which of the severity classes had significantly different volumes foreach of the three volume measures.87The relationship between Ar,nillaria ostoyae and Phellinus weirliThe relationship between A. ostoyae and P. weirii was examined using a Chi-squaretest for dependency. Correlation analysis was also used to compare the severity ratings forboth diseases in the ICR and IDF zones.3.3 Results3.3.1 Individual Tree ResultsHunter’s RangeA total of 3473 trees within 202 plots were examined. Of these trees, 2580 were in theICR zone, while 893 were in the IDF zone. Table 26 summarizes the species distributionwithin each zone by disease status and stocking level. Western redcedar was the only speciesthat had many class 2 infections. Sixteen of the 17, class 2 infected trees examined in the IDF,were western redcedar; while 196 of the 215 class-2-infected trees examined in the ICH werewestern redcedar. Class 2 infections were those that appeared to be held in cheek by the hosttree, as evidenced by callus margins surrounding the basal lesions. The other species had few,if any, class 2 infections. For this reason class 2 and 3 infections were combined in Table 26,which contains the results of the comparisons of the proportion of trees infected amongspecies.Stocking levels were considerably different between the two zones. The ICH zonecontained 921 trees/ha while the IDF zone contained 720 trees/ha. The predominance of manysmall western redcedar understory trees in the ICH zone is probably the single most importantfactor responsible for this difference.88Percentage of total trees infected with Armillariaby species by zone (Hunter’s Range)ostoyae in percentAt Cw Ep Fd Hw Lw P1 Pw Se B! TOTALIDF ZONECLASS 2&3 0 15.2* 1.0 39* 0 12.5 20.3* 0 0 0 9.4(1) (2) (1) (1) (1) (2) (2) (1,2) (1,2) (1,2)TOTAL TREES 3 138 98 407 4 40 202 0 1 0 893TREES/HA 2.4 111.3 79.0 328.1 3.2 32.2 162.9 0.0 0.8 0.0 720ICH ZONECLASS 2&3 0 26.3* 1.7 12.1* 14.2 5.4 33.8* 33.3 8.6 27.8 17.8(1) (4) (1) (3) (3) (2) (4) (4) (2,3) (4)TOTAL TREES 11 1107 234 588 239 203 71 21 70 36 2580TREES/HA 3.9 395.4 83.6 210.0 85.4 72.5 25.4 7.5 25 12.9 921Chi-spuare - 6.013 0.238 18.265 0.569 2.645 4.031 0.000 0.082 0.000 29.807* indicates significant differences (CM-square critical = 3.841) in susceptibility within species between zones.Tree species infection percentages with the same number in brackets beneath them are not significantly differentwithin zones (a=O.05).Larch HillsA total of 760 trees in 32 plots were examined in mature stands within the Larch Hills.Of these, 453 trees were in the IDF zone and 307 were in the ICH zone (Table 27).Phellinus pini was included in the results because the fungus was associated with aconsiderable amount of damage in western larch in both the ICH and IDF. Western larch treeswere often killed (19 of 32 infected) as a result of stems failing due to decay caused by P.pini. Complete records of P. pini were not kept for Hunter’s Range; however, the funguswas often associated with western larch in this area as well.TABLE 26.89TABLE 27. Percentage of total trees infected with Annillaria ostoyae compared amongfive major species along with the percent of trees infected with otherdiseases in percent by species by zone (Larch Hills)Cw P1 Lw Fd EpIDF: A. ostoyae 30.4 15.3 23.1 10.0 7.0(3) (2)(1) (3)(2) (1) (1)P. weirii 0 11.9 7.7 14.6 0P.pini 0 0 23.1 0 0Dead Symptomless 1.0 8.5 13.5 0.8 1.8TOTAL TREES 102 59 104 130 57ICH: A. ostoyae 69.4 44.4 23.6 12.6 14.7(3) (2) (2)(1) (1) (1)P. weirii 0 0 18.2 26.4 0P. pini 0 0 14.5 0 0Dead Symptomless 0 16.7 9.1 5.7 5.9TOTAL TREES 108 18 55 87 34Armillaria ostoyae infection percentages with the same number in brackets beneath them are not significantlydifferent within zones (CM-square critical = 3.841 cr=0.0S).Hunter’s Range and Larch Hills combinedThe proportion of western larch infected did not differ significantly between the twozones in either study area. There was, however, a considerably greater proportion of westernlarch infected in the two zones combined in the Larch Hills (23.3%) than in the Hunter’sRange (6.6%). In the Larch Hills plots, the proportion of western larch infected in the IDFzone was significantly greater than that of Douglas-fir. This result was consistent with thatfound in the Hunter’s Range plots. In the ICR zone, western larch continued to besignificantly more heavily infected than Douglas-fir. The opposite had been found the yearbefore in Hunter’s Range. Another difference between the two areas concerning differences inrelative incidence of A. ostoyae infections existed between lodgepole pine and western larch.The proportion of western larch infected in Hunter’s Range was significantly less than that forlodgepole pine in both the ICH and IDF zones. In the Larch Hills, however, the opposite wasfound with western larch being significantly more heavily infected.90One of the more important differences between the two study areas was in theproportion of trees infected with A. ostoyae. The Larch Hills had a significantly greaterproportion of trees infected (25.6%) than the Hunter’s Range study area (15.9%). The resultfrom the Chi-square test was 40.00 versus a critical value of 3.841 (a =0.05). This differencewas examined further by comparing the two study areas based on the proportion of treesinfected in each of the species individually (Table 28).TABLE 28. Comparison of the percentage of total trees infected with Armillariaostoyae in each of the major species between the Larch Hills andHunter’s Range study areasCw Fd Lw P1 EpLarch Hills 50.5 11.1 23.3 22.1 9.9Hunter’s Range 25.1 8.7 6.6 23.8 1.5Chi-square 57.68 1.075 23.30 0.094 15.79Critical Chi-square a=0.05 is 3.841, a=0.01 is 6.635Not all species were more heavily infected in the Larch Hills than in Hunter’s Range.Those species that did differ significantly in their proportion of trees infected between the twostudy areas had significantly more infection in the Larch Hills than in Hunter’s Range. It isinteresting to note that Douglas-fir, which was one of the three most prevalent species, did notdiffer significantly in the proportion of trees infected between the two areas.Differences in species composition were also compared between the two study areas(Table 29). This comparison was made in order to examine the possibility that the differencesin the proportion of trees infected between the two areas were due to differences in speciescomposition.Some of the minor species from the Hunter’s Range site were not represented in theLarch Hills site, possibly due to the small number of plots on the latter site. The total numberof trees for the Hunter’s range site in Table 29 did include these species, thus, the percentagesfound in this table do not add up to 100. The results in Table 29 indicate that there were91significant differences in the species composition between the two study areas. The mostsignificant difference between the two was due to the high proportion of western larch in theLarch Hills study area.TABLE 29. Comparison of the proportions of the five major species inpercent between the Larch Hills and Hunter’s Range____________________ ___£1.___ _____21.1 10.2 12.17.0 7.9 9.6Cw Fd27.9 28.835.8 28.617.86 0.008 141.6 4.81 4.61Larch HillsHunter s RangeChi-srnisir€Critical Chi-square a=0.05 is 3.841, a=0.01 is 6.635TOTAL7543473The incidences of infection were also compared in four of the species among the fourdifferent areas simultaneously. The proportion infected of three species (western redcedar,western larch and lodgepole pine) were compared to the Douglas-fir standard among fourareas; Hunter’s Range ICH, Hunter’s Range ICH, Larch Hills ICR, Larch Hills IDF (Table30). The methods used in this analysis were the same as those used in Chapter 2 to rank thespecies of conifer regeneration.TABLE 30. Ranking of conifer species compared to a Douglas-fir standard onthe basis of the incidence of Armillaria ostoyae infection in maturestands (approx. 120 years) within the ICH and IDF zones of Hunter’sRange and Larch HillsSPECIES #TREES #INFECED K-VALUE CHI-SOUARE CRITICAL CHI-SOUAREFd 1212 111 1.0000 - -Cw 1455 418 2.5662 57.950 5.991Lw 402 53 1.2200 29.377 5.991P1 350 82 3.3221 14.578 5.991The highly significant Chi-square values confirm the differences eluded to earlier. Theproportion of trees infected with A. ostoyae for a given species differ depending on the site inwhich the species are found. Due to the large differences in the proportion of trees infected92among the different sites for the three species above, it is not possible to state a single K-valuefor each species. Thus, unlike the regeneration results in Chapter 2, a quantitative ranking ofspecies based on the incidence of infection can not be made with the mature tree data. Despitethe significant differences in disease incidence among sites, the K-values in Table 30 indicatesome important results. All three of the species that were compared to Douglas-fir had K-values higher than 1.000. Thus, over the four areas (ICH and IDF in both Hunter’s Rangeand Larch Hills) western redcedar, lodgepole pine and western larch were generally moreoften infected than Douglas-fir.Incidence of Disease among Site Units (Hunter’s Range)Differences in the incidence of A. ostoyae infections within each species among siteunits within each zone were examined (Table 31). Three of the species (western white pine,Engelmann spruce, and subalpine fir) were combined due to the low numbers of trees in thesespecies in the ICH zone. Only those species that were present in sufficient numbers to testwith a CM-square test were included in the analyses.TABLE 31A. Percentage of trees infected in percent for each species in each site unitin the ICR zoneSITE UNIT Fd Cw Hw Lw (Pw,Se,Bl)% total % total % total % total % totalICHmklOl 11.4 35 15.4 91 0.0 5 0.0 19 17.5 57ICHmw2Ol 31.6 19 34.3 102 21.7 60 0.0 3 50.0 8ICHmw2Olys 11.9 227 28.0 436 3.6 55 10.8 74 8.3 12ICHmw2O2 12.0 83 0.0 19 - - 0.0 13 - -ICHmw2O3 9.3 182 28.6 360 13.2 38 3.4 89 42.9 7ICHmw2O5 16.7 42 17.2 99 17.3 81 0.0 5 11.6 43CM-square 7.866 18.125 6.420 4.345 23.56Critical 11.070 11.070 7.815 5.991 9.48893TABLE 31B. Percentage of trees infected in percent for each species in each siteunit in the IDF zoneSITE UNIT Fd Cw Lw P1% total % total % total % totalIDFmwlOl 4.5 247 16.4 128 15.4 130 26.2 26IDFmwlO4 3.9 128 0.0 6 0.0 6 11.5 61IDFmw2Ol 0.0 32 0.0 4 12.5 8 0.0 11Chi-square 1.429 1.641 0.588 4.815Critical 5.991 3.841 3.841 5.991Western redcedar was the only individual species that had significantly differentincidence of infection among biogeoclimatic site units within either zone. Subalpine fir,western white pine and Engelmann spruce combined also had significantly differentproportions among the site units in the ICR. The proportion of western redcedar infected inthe ICHmw2O2, ICHmw2O5 and ICHmklOl site units was significantly lower than that in theICHmw2Ol, ICHmw2Olys and ICHmw2O3 site units. There were not enough westernredcedar present in the IDF zone to make comparisons between site units.There were no significant differences among site units for Douglas-fir or western larchin either the IDF or the ICH zones. Western hemlock did not appear in the TDF. Within theICH zone, there were no significant differences between site units in the proportion of westernhemlock trees infected with A. ostoyae.Comparison of Methods of Data CollectionA follow-up study was conducted on the Larch Hills in order to determine moreaccurately the number of infected trees in mature stands within the ICHmw2 (Table 32). Thisstudy compared two methods of root disease data collection and the number of A. ostoyaeinfections found with each method. The 1992 method was used on the Hunter’s Range site(202 plots) and the northwest Larch Hills site (32 plots). The 1993 method was used in theplantation study (Chapter 2). The 1993 method involved the removal of the litter layer from94the base of the tree. The data from these plots were also used to verify the amounts ofinfection found in the stump examinations performed in Chapter 2.TABLE 32. Comparison of the 1992 and 1993 methods of data collectionregarding the number of infections found, and the percent of totaltrees missed using the 1992 methodPLOT TOTAL NO. INFECTED NO. INFECTED PERCENTCONIFERS 1992 1993 MISSED1 29 10 16 20.72 25 14 18 16.03 22 10 14 18.24 28 13 18 17.95 24 8 10 8.36 22 5 7 9.17 21 14 18 19.08 19 11 14 15.89 21 10 11 4.810 19 4 10 31.6TOTAL 230 99 136 16.1Of the 37 trees with “missed” infections, 20 were Douglas-fir, 10 western hemlock, 4Engelmann spruce, 2 western redcedar, and 1 western larch. Four of the missed Douglas-firinfections were callused-over lesions.Although the more intensive methods detected 16.1 % more infected trees, the majorityof these trees were infected with class 2 infections. Thus, the average plot severity ratingincreased only slightly from 11.1 to 11.5. Using the 1992 methods, the Larch Hills study areahad far more infections than the study area on Hunter’s Range with 43.0% of the treesinfected, compared to only 15.6%. The reasons for this difference are not clear. It is possiblethat this area in the Larch Hills area was simply more heavily infected than Hunter’s Range.953.3.2 Plot Results (Hunter’s Range)A summary of the plot results including the number of plots/site unit, number ofdisturbed plots and the number of diseased plots is found in Table 33. There was an unequalnumber of plots between the two zones. Within the ICH zone 140 plots were examined while62 plots were examined within the IDF zone. The proportion of plots in each zone may beused as a rough approximation of the distribution of zones within the study area. Using thisapproximation 140/202 or 69.3% of the area was in the ICH biogeoclimatic zone while 62/202or 30.7% was in the IDF.The majority of the 202 plots (86.1 %) were diseased with either A. ostoyae or P. weiriior both. Armillaria ostoyae was the most prominent forest pathogen, with over 76.7% of theplots exhibiting evidence of infection. Phellinus weirii was present on 18.8% of the plots.Disturbance was also quite widespread throughout the study area with 46.5% of the plotshaving evidence of logging either in or just outside the plot perimeter. Most of the disturbedplots (86.2%) exhibited evidence of root disease. Similarly, 85.2% of the undisturbed plotsalso contained evidence of root disease.TABLE 33. Plot distribution by site unit, disease status, and disturbance historyICH IDFSITE UNIT mw2Ol mw2Olysmw2O2 mw203 mw2OS mklOl mwlOl mwlO4 mw2Ol TOTAL#of PLOTS 14 41 10 44 16 15 41 13 8 202% DISTURBED 78.6 48.8 20.0 50.0 75.0 66.7 41.5 0.0 0.0 94% ARMILLARIA 92.9 85.4 70.0 86.4 81.2 86.7 63.4 61.5 25.0 155% PHELLINUS 0.0 12.2 0.0 34.1 6.2 0.0 17.1 30.8 75.0 38% ARM & PHEL 0.0 9.8 0.0 22.7 6.2 0.0 4.9 7.7 12.5 19% of DISTURBEDPLOTS with ARM. 90.9 85.0 50.0 100.0 83.3 90.0 70.6 0.0 0.0 81% of UNDISTURBEDPLOTS with ARM. 100.0 85.7 75.0 95.5 75.0 80.0 79.2 84.6 87.5 92Correlation analysis was used to determine which of the factors examined (i.e,disturbance, site index, elevation Phellinus severity) was responsible for the greatest amount ofvariation in A. ostoyae severity (the dependent variable). Prior logging disturbance was the96most closely correlated variable out of those analyzed (r = 0.329). The correlationcoefficients for the other factors were: site index -0.0 16, elevation 0.040, and Phellinusseverity -0.168. The critical correlation coefficient for n=200, and a=0.05 is ± 0.138.Armillaria ostovae severity vs DisturbanceArmillaria ostoyae severity was significantly higher in plots within both the ICH andthe IDF zones when these plots also contained evidence of prior logging activity (Fig. 4).This relationship was significant in both the combined zone analysis and the individualanalyses for the ICH and IDF zones (Table 34). The relationship was also significant for bothof the zones when two disturbed classes were included with the undisturbed class. Asmentioned earlier (page 83) there were two disturbed classes. The difference between the twowas based on whether or not the evidence of disturbance was in or out of the plot.ARMILLARIA SEVERITY vs DISTURBANCE(ICH & IDE combined)w>wC/)-J(F (F ISTUMPS IN PLOT STUMPS OUTSIDE PLOT NO STUMPSDISTURBANCE EVIDENCEFIGURE 4. Armillaria ostoyae severity compared between disturbed and undisturbedplots (ICR and IDF combined)97TABLE 34. ANOVA results for Annillaria ostoyae severity vs disturbance classin the ICR and IDF zonesSOURCE SUM OF SQUARES DF MEAN-SQUARE F-RATIO pBETWEEN GROUPS 723.219 2 361.609 14.544 0.000WI1’HIN GROUPS 4947.811 199 24. 863Prior disturbance did not influence the presence of A. ostoyae on a site. Plots thatcontained evidence of prior disturbance were no more likely to exhibit signs of A. ostoyaeinfection than those plots that were not disturbed. Of the 94 disturbed plots from the entirestudy area, 81 were also diseased (86.2%). Of the 108 undisturbed plots, 92 were diseased(85.2%).Annillaria ostovae Severity vs ElevationThe correlation between Armillaria ostoyae severity and plot elevation data wasdetermined to assess whether elevation within the ICH and IDF zones influenced thedistribution and impact of the disease (Fig. 5). The plot elevations in the study area rangedfrom 410m to 144Gm. There was no significant relationship between A. ostoyae severity inplots and their elevation (r=0.040, n=202, p =0.596).Elevation did influence the distribution and impact of P. weirii. This relationship wassignificant in both the ICH and the IDF zones. There was significantly more P. weirii in theplots at lower elevations than those higher up the slope (r=-0. 197, n=202, p=O.OOS). Thepossibility that logging disturbance was confounding the results tested above was examined. Aone-way ANOVA was used to test whether or not the mean elevation of the undisturbed plotswas significantly different from that of the disturbed plots. There was no significant differencebetween the two. The mean elevation of the disturbed plots was 849.7m while that of theundisturbed plots was 831.7m.98ARMILLARIA SEVERITY vs ELEVATION(ICH & IDF combined)1600* w1400* * )IE120O******100o * *I— I )IE0 *-w *_*_J ni-v . * * 0 * *w ovi. 0 * ** * ** ** *400o***25ARMILLARIA SEVERITYFIGURE 5. The relationship between Armillaria ostoyae severity and plot elevation inthe ICII and IDF zonesArmillaria ostovae Severity vs Site IndexThere was no significant relationship between A. ostoyae severity and site index usingthe site indices for interior Douglas-fir (r=-0.016, n=202, p =0.824).Armillaria ostovae Severity vs Biogeoclimatic Site UnitThere were no significant differences in A. ostoyae severity among site units withineither of the zones (Fig. 5). A one-way analysis of variance of A. ostoyae severity versus siteunit within both zones proved to be insignificant (Table 35). The A. ostoyae severity wassignificantly different between the two zones (1’-test statistic=-3. 150, p =0.002) beingsignificantly higher in the ICH than in the IDF.99TABLE 35. ANOVA results for Armillaria ostoyae severity vs biogeoclimatic site unit inthe ICH and IDF zonesSOURCE SUM OF SOUARES DF MEAN-SOUARE F-RATIO pICH zoneBETWEEN GROUPS 67.601 5 13.520 0.507 0.770WITHIN GROUPS 3572.271 134 26.659IDF zoneBETWEEN GROUPS 104.722 2 52.361 1.863 0.164WITHIN GROUPS 1658.389 59 28.1089&7.Lii>&wCl)z&I.ARMILLARIA SEVERITY vs SITE UNITICH AND IDE ZONESFIGURE 6. The relationship between Armillaria ostoyae severity and biogeodilmatic siteunit in the ICH and IDF zonesThe individual tree results described earlier indicated that only western redeedar hadsignificantly different proportions of trees infected between site units. Thus, it is notsurprising that the combined species analysis did not indicate any significant differences in A.ostoyae severity among site units.HMW2OIY HMW2O5 • HMW23 • HMW2O2 • HMW2OI RMKIOI • FMW2OI • FMWIO4 • FMWIO1BIOGEOCLIMATIC SITE UNIT100Armillaria ostovae Severity vs Timber VolumeThere were significant differences in mean conifer volume among the four A. ostoyaeseverity classes (page 88) for the ICH and IDF zones combined (Table 36). Plots withseverity class 15 had significantly lower mean conifer volumes than those in the other threeclasses.TABLE 36. ANOVA results for Annillaria ostoyae severity vs conifer volumein the ICH and IDF zones individually and combinedSOURCE SUM OF SQUARES DF MEAN-SQUARE F-RATIOICH and IDFBETWEEN GROUPS 431.302 3 143.767 7.378 0.000WITHIN GROUPS 3858.261 198 19.486ICHBETWEEN GROUPS 348.290 3 116.097 5.852 0.001WITHIN GROUPS 2698.169 136 19.839IDFBETWEEN GROUPS 111.611 3 37.204 2.123 0.107WITHIN GROUPS 1016.591 58 17.527There was significantly less conifer volume in A. ostoyae severity class 15 than in anyof the other severity classes in the ICH zone (Table 37). There were no significant differencesamong the A. ostoyae severity classes in the IDF zone.101TABLE 37. The relationship between Armillaria ostoyae severity and conifervolume with the number of piots in eacI severity class and theassociated mean conifer volume/plot (m /ha) for the ICH and IDFzonesICH ZONE IDF ZONEArm. Severity Class #Plots Conifer m3/ha #Plots Conifer m2/ha0 21 510 26 3635 53 508 15 45910 48 437 15 36115 18 266 6 208Total Plots 140 62The relationship between birch volume and A. ostoyae severity was significant for boththe IDF zone and the ICH zone (Table 38).TABLE 38. ANOVA results for Armillaria ostoyae severity vs paper birch volume inthe ICH and IDF zonesSOURCE SUM OF SOUARES DF MEAN-SOUARE F-RATIO pICH zoneBETWEEN GROUPS 7.821 3 2.607 3.122 0.028WITHIN GROUPS 113.571 136 0.835IDF zoneBETWEEN GROUPS 3.224 3 1.075 3.826 0.014WITHIN GROUPS 16.294 58 0.281Paper birch volume was significantly higher in those plots in A. ostoyae severity class15 in the combined zone analysis. For both of the zones individually the only significantdifferences in paper birch volume among the four severity classes was between Class 5 andClass 15 (Table 39).Differences in total volume among the four A. ostoyae severity classes were also tested.With both zones combined the ANOVA was significant. There was significantly less totalvolume in the most severely infected class than in the other three classes. Significantdifference in total volume were found in the ICH zone was among severity classes 15, 0 and 5.Severity class 15 had significantly lower volume than class 0 and class 5 but not class 10.102There was no significant difference in total volume among the four severity classes in the IDFzone.TABLE 39. The relationship between Armillaria ostoyae severity and birchvolume with the number of plots in each severity class and theassociated mean birch volume/plot (m3/ha) for the ICH and IDFzonesICH ZONE IDF ZONEArm. Severity #Plots Birch m3lha #Plots Birch m0 21 20.0 26 14.55 53 21.0 15 4.010 48 27.7 15 26.715 18 57.2 6 42.7Total Plots 140 62The Relationship between Armillaria ostovae and Phellinus weirli in the ICH and IDFBiogeodimatic zonesThe IDF zone had significantly more plots infected with P. weirli than the ICHzone. Thus, it appears that this disease possibly prefers drier sites found in the IDF. Therelationship between A. ostoyae and P. weirii is not clear. Results from the first field seasonindicated that the two diseases are not as compatible on the same site as once believed. Thecorrelation analysis of the two disease severity ratings from the two zones combined wassignificant (r=-0. 170, n=202 p =0.016). Within the ICH zone, the occurrence of the twodiseases appeared to be independent of each other. The two diseases did not occur togetherany more often than would be expected completely randomly. Within the IDF zone, P. weiriioccurred together with A. ostoyae less often than would be expected if they were randomlydistributed (Table 40).103TABLE 40. Test of independence between Armillaria ostoyae and Phellinus weiriiin the ICR and IDF zonesICH ZONE: ARMILLARIA NO ARMILLARIA TOTALPHELLINUS 15 (17.85) 6(3.15) 21NOPHELLINUS 104(101.15) 15 (17.85) 119TOTAL 119 21 140Chi-square= 3.569, Critical Chi-square= 3.841IDF ZONE: ARMILLARIA NO ARMILLARIA TOTALPHELLINUS 4 (9.9) 13 (7.1) 17NO PHELLINUS 32 (26.1) 13 (18.9) 45TOTAL 36 26 62Chi-square= 11.59, Critical Chi-square= 3.841(expected values are bracketed)These relationships could not be tested using the data collected on Larch Hills. Out ofthe 32 plots examined, P. weirii occurred in 12 and A. ostoyae occurred in all but two of theplots. All of the plots that contained evidence of P. weirli also contained evidence of A.ostoyae. Since nearly 100% of the plots were infected with A. ostoyae it was not possible totest for independence between the two diseases.3.4 Discussion3.4.1 The Ranking of Conifer Species Based on Incidence of Armillaria ostoyaeInfections in the ICR and IDF ZonesMorrison’s (1981) ranking of tree species susceptibility was based on trees that wereknown to have been infected. A susceptible species was one that once infected was soonkilled. Resistant species were rarely killed when infected. Western larch has been consideredone of the most resistant conifer species within the southern interior of B.C.(Morrison 1981).Hagle and Goheen (1988) also considered western larch to be one of the most A. ostoyaeresistant species based on their observations in mature stands in the Intermountain Northwestarea of the United States. The question of western larch being more resistant than other104conifer species was reviewed in this research. The methods used in this study differedconsiderably from those of Morrison (1981). In this study, the species were ranked based onlyon the presence of above ground symptoms of A. ostoyae. It was not known whether or not asymptomless tree had been infected by the disease. Work by Morrison in the ICH suggeststhat the proportion of conifer trees infected in this zone ranges from 60-100%, based on belowground evidence (pers. corn.). Thus, the probability that the symptomless trees were notchallenged by A. ostoyae is low. Despite differences in methods, the ranking of certainconifer species as more “susceptible” than others in this study differed considerably from thosefound by Morrison (1981) listed on page 9. The proportions of trees infected for three specieswere compared to a Douglas-fir standard measure of disease incidence (Table 30). Theproportion infected for Douglas-fir, western redcedar, western larch and lodgepole pine variedbetween the ICR and IDF zones on both the Hunter’s Range and Larch Hills sites. Theproportion infected for the three species varied significantly from the Douglas-fir standard.This variation among the four sites made it impossible to state definitively which species wasproportionately more frequently infected. Whether a species was more “susceptible” thananother appeared to depend more on site than on species characteristics.The K-values (Table 30) represented a quantitative ranking of disease incidence. Sincethe proportion infected for each species varied significantly from the standard, the confidencelevel in the K-values is low. Despite this, the K-values still indicated differences in A. ostoyaeinfection incidence among the four major species. Douglas-fir had the lowest relative rate ofinfection of the four species. Lodgepole pine and western redcedar had considerably higherrates of infection. Western larch was also more heavily infected than Douglas-fir though notto the same degree as lodgepole pine and western redcedar.The fact that the relative proportions of trees infected, for most conifer species, variedfrom site to site was made clear by examining the two zones within each study areaindividually. Within the ICH zone on Hunter’s Range the proportion of western larch infectedwith A. ostoyae was significantly less than that of any of the other conifer species (Table 26).However, the proportion of western larch infected was significantly greater than the proportion105of Douglas-fir infected in the IDF zone in that study area. The results from the Larch Hillsindicated that the proportion of western larch infected with A. ostoyae was not significantlyless than that of any of the conifer species in both the IDF and ICH zones. The proportion ofDouglas-fir and lodgepole pine infected in the Larch Hills was considerably less than thatfound in Hunter’s Range, while the opposite was found for western larch.The species ranking based on incidence of A. ostoyae infection in the Larch Hills(Table 27) differed from that determined in the Hunter’s Range data (Table 26). The reasonsfor this could lie in the difference in species composition between the two sites. There weresignificantly more western larch trees on the Larch Hills site and significantly fewer westernredcedar. It is possible that the reason the proportion of larch trees infected was significantlygreater than that found on Hunter’s Range was due to the higher stocking of that species onLarch Hills. When a species is relatively abundant on a site, it is perhaps easier for disease tospread within that species than to spread between individuals of differing species. Trees of thesame species would have the same root form and would tend to graft roots betweenindividuals. Trees of different species that were adjacent to each other, would be less likely tohave roots in contact due to differences in root system form. The proportion of total stockingconsisting of Douglas-fir did not differ significantly between the two sites, nor did theproportion of that species infected. The same was true for lodgepole pine. However, forwestern larch, the proportion of total stocking on the Larch Hills site was significantly greaterthan that on Hunter’s Range and the proportion of trees infected was also significantly greater.More work is required to determine definitively whether or not western larch is indeedmore resistant than Douglas-fir and other conifer species. More work is required in generalregarding the ranking of species susceptibility to A. ostoyae in the southern interior of BritishColumbia. The Larch Hills would appear to provide an excellent study area to test for therelative differences in species susceptibility to A. ostoyae. Hadfield (1984) stated that it isdangerous to make generalizations about the susceptibility of tree species to Armillaria rootrot. It is quite apparent from the results of this research that his statement was quite sound.1063.4.2 The Influence of Prior Disturbance on the Expression of Armillaria ostoyae in theICH and IDF zonesPast logging disturbance was closely associated with high A. ostoyae severity in plots inboth the ICH and the IDF zones. The correlation between prior disturbance and A. ostoyaeseverity had the highest r-value of any of relationships for any of the site factors examined inthis study. Differentiating between the evidence of disturbance inside the plot and that outsidedid not significantly affect the strength of the relationship (Fig. 4). Both measures ofdisturbance had very similar effects on the severity of A. ostoyo.e within the plots. There weresignificantly higher A. ostoyae severity ratings on those plots that exhibited evidence of pastlogging disturbance than those that were not disturbed.The past logging disturbances did not affect the area of forest land infected with A.ostoyae. There was a roughly equal proportion of plots infected in those areas that had noevidence of disturbance as those that did. Of the plots that were disturbed, 86.3% were alsodiseased, while 85.3% of the undisturbed plots were diseased. However, the A. ostoyaeseverity rating of the disturbed plots did differ significantly from that for the undisturbed plots.Thus, the proportion of trees infected in the disturbed areas was significantly higher than theproportion of trees infected in the undisturbed areas. These results agree with those of Shaw etal. (1976) in their study of A. mellea in south central Washington. They found that the area ofA. mellea infection following cutting remained essentially constant, but the volume losses morethan doubled. The effects of cutting history on the incidence of root disease that Byler etal.(1987) observed in the Crow Creek compartment, Lob National Forest, Montana differedfrom those found in Hunter’s Range. They found that cutting history increased both theseverity of infections and the incidence of root disease. Ten percent of the stands in theirstudy had some evidence of past disturbance. In those stands, the proportion of plots with rootdisease damage was roughly double that of plots that had no evidence of past cutting. Thesedifferences aside, there is certainly evidence to suggest that partial cutting in mature standsincreases the probability of losses to A. ostoyae.1073.4.3 The Influence of Elevation on Root Disease ExpressionThe influence of elevation on root disease activity has been examined in several studiesin the Intermountain Northwest region of the United States. Hobbs and Partridge (1979)examined the incidence of root rots along an elevation gradient in mixed conifer stands innorthern Idaho. They reported that A. mellea was not affected by elevation, describing itsoccurrence as ubiquitous. They found that P. weirii was affected to some degree by elevationand that this root disease was generally found below 1500m elevation in that region. TheHunter’s Range results agree with those found by Hobbs and Partridge (1979). There was nosignificant relationship between the elevation of a plot and the incidence or severity of A.ostoyae infection. Phellinus weirli incidence was dependent on elevation. There wassignificantly more P. weirii infections at lower elevations. Williams and Marsden (1982)found that the probability of root disease center occurrence was inversely related to elevation.Their study did not differentiate between A. ostoyae and P. weirii; thus, their conclusions werenot necessarily contrary to those found in this study.3.4.4 The Relationship between Armillaria ostoyae Distribution and the BiogeodilmaticSite Classification in the Southern Interior of B.C.Several studies in the Intermountain Northwest region of the United States have relatedboth incidence and severity of root disease to the habitat classification system used in that area(Byler et a!. 1987, Hagle 1985, McDonald et a!. 1987). None of the literature regarding A.ostoyae in the southern interior of B.C. has connected the incidence and severity of the diseasewith the BEC system at any finer detail than the zonal level. This research was designed todescribe differences in the role of A. ostoyae within two of the zones on a more detailed scale.The results of this study indicated there were no significant differences in the incidenceand severity of A. ostoyae among site units in either zone. The only individual tree species inwhich the site unit appeared to influence the proportion of trees infected was western redcedar.108The proportions of Douglas-fir, lodgepole pine, western larch and western hemlock that wereinfected did not differ significantly among site units.The lack of significant differences in A. ostoyae incidence among site units wasprobably due, in part, to the manner in which the site units were identified at each plot. Thedifference between some of the site units was sometimes decided by the presence or absence ofa single plant species. However, the differences between the driest and wettest site unitswithin a zone would have been obvious.The proportion of western redcedar trees that were infected was significantly less in themore moist site units than in the drier ones within the ICH. There was, however, asignificantly greater proportion of western redcedar infected in the ICH than in the drier IDFzone. Thus, greater moisture did not appear to reduce the probability of western redcedarbecoming infected on the larger zonal scale. This apparent discrepancy could be explained onthe basis of differing levels of inoculum between the two zones. The ICH zone simply hadmore A. ostoyae inoculum than the IDF zone; thus, any trees within the ICH were at a greaterrisk of becoming infected. If both zones had the same levels of inoculum, then perhapswestern redcedar would have had a lower proportion of trees infected in the ICH than in theIDF zone.There were significant differences in the incidence and severity of A. ostoyae betweenthe two zones based on the Hunter’s Range data. The ICH zone had a significantly greaterproportion of diseased plots than the IDF zone. These results agree with those of Byler etal.(1987) and Hagle (1985) who both found that the more productive habitat types suffered themost severe damage. These results are in contrast to those found by McDonald et al.(1987)who stated that the most productive sites in National Forests of the inland western UnitedStates showed low incidence of A. ostoyae. McDonald et al.(1987) stated that the incidenceof the disease showed a strong tendency to decrease as stand productivity increased. It isapparent that the latter argument is not the case in the southern interior of B.C. The ICH zoneis after all the most productive zone in the interior of B.C. (Meidinger and Pojar 1991).1093.4.5 The Relationship Between Armillaria ostoyae Severity and StandProductivity in Terms of Timber VolumeThere has been very little research done on the impacts of A. ostoyae on forestproductivity in terms of timber volume in the southern interior of B.C.. The ICH zone is thesecond most productive forest zone in Canada, yet it quite possibly has the highest incidence ofA. ostoyae of any forest zone in the country. This study indicates that A. ostoyae is widespread in some areas of the zone. It is clear that A. ostoyae cannot be killing all trees or theICR would not be nearly as productive as it is. The results of this study indicate that therewas a significant relationship between A. ostoyae severity class and conifer volume in the ICHzone and the ICR and IDF zones combined. There was no significant relationship betweenconifer volume and A. ostoyae severity class in the IDF zone. However, the most severelyinfected plots in the IDF had considerably less conifer volume than the less infected plots (208m3/ha vs 459 m3/ha, Table 37).Whether these relationships between A. ostoyae severity class and conifer volume weresignificant or not depended in part on how the severity classes were defined. The methodsused in this study combined plot severity ratings into four classes (page 88). Combining theplot severity ratings into a greater number of smaller classes could yield different results. Thesame could be said for fewer large classes. Regardless of how the severity classes are defined,it is apparent that the most severely infected stands in both zones have considerably lessconifer volume than less severely infected stands. It is also clear that A. ostoyae can be quiteprevalent in a stand without having a significant negative impact on conifer volume (Table 37).In the ICH zone conifer volumes did not drop significantly until A. ostoyae severity class 15.Thus in this zone, perhaps as many as 40% of the stems could have shown signs of infectionwith as many as 30% having Class 3 infections, before volumes decreased.There was also a significant relationship between paper birch volume and A. ostoyaeseverity class in the ICH and IDF zones combined. Plots with a severity rating greater than14.0 had significantly greater birch volumes than those plots with lower A. ostoyae severityratings (Table 39). The fact that there was a significant relationship was in part an artifact of110the methods used to determine the A. ostoyae severity rating. The rating was based on theproportion of conifer trees infected out of the total number of conifers. In those plots with asmall conifer component and a large birch component a single infected conifer translated into ahigh proportion of conifers infected. Thus, such areas with high A. ostoyae severity ratingswould quite likely have a large birch component. This is probably not far from the truth in themajority of the stands in the ICH. Often those areas most heavily infected with A. ostoyae areconverted from conifer to deciduous forest types. If the severity ratings were based on theproportion of trees infected out of the total trees including birch, the ratings wouldunderestimate the effects of A. ostoyae in these stands. The severity rating used in this studyincorporates the effects of severe A. ostoyae infestations on species composition.3.4.6 The Relationship Between Phellinus weirli and Armillaria ostoyaeThe results regarding the relationship between A. ostoyae and P. weirli in this researchreflect the differences in opinion among the various workers who have examined this subject.The results from the original study on Hunter’s Range are reported in Table 40. They indicatethat the occurrence of the two diseases was independent of each other in the ICH zone, andwas significantly negatively related in the IDF zone. This is contrary to the results of Filipand Goheen (1982) who found that the two diseases were often found in association with eachother. The ICR results, in part, support the claim of Hansen and Goheen (1989) who attributethe association between the two diseases to chance and to primary-secondary relationships.Primary-secondary relationships refer to situations where a host is infected by a pathogen, suchas P. weirli, but not killed. The vigor of the infected host is reduced making it morevulnerable to A. ostoyae. Such relationships would result in more frequent associationbetween the two diseases than random. Attributing the relationship between the two disease tochance is probably most appropriate for the ICH. The negative relationship between A.ostoyae and P. weirli in the IDF zone requires more study.111In the Larch Hills P. weirli was only found in association with A. ostoyae. None of theP. weirli infected plots contained evidence of only that disease. All P. weirli plots alsocontained evidence of A. ostoyae infections. This result is hardly surprising since 30 out of 32plots exhibited evidence of A. ostoyae infections.3.4.7 The Value of Surveying for Armillaria ostoyae in Mature Stands in the ICHand IDF zonesThe follow-up study concerning root disease survey methods revealed that an averageof 16.1 % of the infections in that study would not have been identified with the methods usedin the Hunter’s Range study the previous year. These results have important implications forforest managers who are required by regulations in this province, to carry out root diseasesurveys in areas believed to be infected. Clearly, the amount of root disease detected in astand has a great deal to do with the methods used to survey that stand. The effort required tofind the additional 16.1% of A. ostoyae infections was considerable. In order to find all ofthese infections the bark had to be removed from the lateral root up to a distance of 30cm fromthe root collar. Had the roots been excavated even further, the amount of infection in thesestands would no doubt have been higher. The question then becomes: “What is theappropriate degree of accuracy required in a root disease survey?”. From the forest manager’sviewpoint, does it matter if a root disease survey has missed 16% of the actual number ofinfections in a stand?It is quite clear that A. ostoyae is present in the majority of sites within the transitionbetween the ICH and IDF biogeoclimatic zones. Thus, it appears the value of surveying forthe presence of A. ostoyae in this area is minimal. The results from the plantation study inChapter 2 indicated that the relationship between A. ostoyae infections in stumps and plantationmortality was poor. Thus, the results from a root disease survey would not significantlyimprove any predictions of future plantation health. These two arguments combined seriouslyquestion the value of A. ostoyae surveys in the ICH and ICH/IDF transition portions of thesouthern interior of B.C.112CHAPTER 4 CONCLUSIONSSeveral important conclusions may be made based on the results of this research. Aword of caution should be included with these conclusions as they are based on a small numberof study areas. With regards to the behavior of A. ostoyae in plantations, six clear conclusionsmay be made. First, the negative impacts of brushing in plantations within the ICH are quiteobvious. Ar,nillaria ostoyae is present in the majority of sites within the Interior CedarHemlock zone. The results from the two 1969 plantations examined in Chapter 2 illustratevery clearly the impact of brushing stands in this zone. The losses to A. ostoyae in theunbrushed plantation were significantly less than those suffered in the brushed plantation.Second, the proximity of infected stumps to trees in 25-year-old plantations does notsignificantly increase the probability of those trees becoming infected. The proximity ofinfected stumps to trees in 10-year-old plantations does influence the probability of those treesbecoming infected. This conclusion is drawn from the comparison of the two different plotsizes used in the study of A. os:oyae in plantations. Using a smaller plot size did not improvethe relationship between the infected stumps and infected trees in the older plantation. Theopposite was found in the younger plantation. This has very important implications for freegrowing surveys and the current policy of the Ministry of Forests in B.C.. This policy statesthat any tree within three meters of a stump infected with A. ostoyae cannot be considered freegrowing. The results of this research indicate that the distance that a tree is from an infectedstump is not strongly related to the probability of that tree becoming infected. Trees in the 25-year-old plantation had survived 25 years in close association with infected stumps.Third, the question of ranking species susceptibility to A. ostoyae requires more studyin both young and mature stands. Whether or not western larch is significantly more resistantto A. ostoyae infection than the other conifer species is not certain. The relationship betweenPhellinus pini and western larch survival in mature (120-year-old) stands also deserves moreresearch. Phellinus pini had a significant impact on western larch survival in the mature standsurveyed on Larch Hills. Before western larch is prescribed on a large scale as the species of113choice in areas infected with A. ostoyae in the ICH, the relationship between this tree speciesand both A. ostoyae and P. pini in mature stands should be studied further.Fourth, the use of locigepole pine in plantation management in the ICH zone should bereviewed. The results from the mature stands indicated that lodgepole pine suffered one of thehighest rates of mortality due to A. ostoyae of any of the conifer species. Douglas-fir wassignificantly more tolerant to this pathogen than lodgepole pine in the mature stands.Fifth, the fact that a mature stand may be quite heavily infected with A. ostoyae doesnot necessarily mean that a future plantation, following harvest on that area will not beproductive. The results from the unbrushed 1969 plantation show that the overall stockinglevel in the 25 year old plantation was the highest of the four plantations although the priorstand was heavily infected with A. ostoyae. Although a fairly large proportion of the stockingwas made up of ingress there was still over 400 stems/ha of the original planted stock. Asmentioned above, any additional entries into that plantation for brushing or spacing wouldplace the health of the young stand at a much greater risk. It should be kept in mind that if theintended stand composition is largely based on a single species, it will not be realized. Thepresence of A. ostoyae on a site ensures species diversity.The sixth conclusion based on the plantation study is concerned with the use of rootdisease surveys for predicting future losses in plantations. The number of infected stumps wasthe stump measure most highly correlated with losses in plantations. The size of the infectedstumps as measured by stump basal area was not as highly correlated. Neither of these stumpmeasures were very strongly correlated. It may be concluded from these results that intensivesurveying for root disease in mature stands prior to harvest may not be useful. The estimatesof future losses in the plantations would simply not be very accurate. Furthermore, theaccuracy of estimates of infection prior to harvest are equally as poor due to the difficultiesassociated with accurately assessing the amount of infection. In order to accurately estimatethe number of infected trees, a considerable amount of effort is required. The costs ofobtaining accurate estimates of A. ostoyae infection in mature stands are not offset bysubstantial gains in the accuracy of predictions of future losses in plantations.114There are four major conclusions to be drawn from the study of A. ostoyae in maturestands in the ICH and the ICH/IDF transition zones. First, as already mentioned, the rankingof species by the incidence of infection differs considerably from the ranking of speciessusceptibility that is most prevalent in the literature. The ranking of species based on theincidence of infection depends more on the relative abundance of that species on a site and thesite itself than on the species. Additional research is required to determine more definitivelythe rankings of conifer species suceptibility to A. ostoyae.Second, the use of the BEC system beyond the zonal scale to predict the incidence andseverity of A. ostoyae in the southern interior is not appropriate. The incidence and severity ofthe disease within site units in the ICHmw2, and IDFmwl variants did not differ significantly.Third, of the site factors recorded for each plot, logging disturbance was most closelyassociated with high A. ostoyae severity. Those plots with evidence of past loggingdisturbance had significantly higher A. ostoyae severity ratings than the undisturbed plots.Fourth, although most of the plots in the ICH zone contained trees that were infectedwith A. ostoyae, the conifer volume of the majority of these plots was not negatively affected.Only in the most severely infected plots did the volume drop significantly. Few plots were inthat category. The fact that only the most severely infected plots had significantly lowerconifer volumes in the ICR, helps to explain some apparent contradictions. The ICH isconsidered one of the most productive ecological zones in Canada, yet A. ostoyae is prevalentthroughout the zone. However, it is not likely that the volumes currently being harvested offmature stands in the ICR will be realized in the future, under the current forest managementstrategies. As pointed out in both the plantation and mature stand studies, disturbance whetherfrom brushing or partial cutting significantly increases the severity of A. ostoyae. If both ofthese treatment options continue to be prescribed, particularly brushing, the productivity of theICH zone will no doubt decline.115BIBLIOGRAPHYAldrich J.H.; Nelson F.D. 1985. Linear Probability, Logit, and Probit Models. SagePublications, series no. 07-045. Beverly Hills and London. 83 p.British Columbia Forest Service, Inventory Division. 1976. Whole stem cubic metre volumeequations. Victoria B.C.Bloomberg, W.J. 1983. A ground survey method for estimating loss caused by Phellinusweirli, IV. Multiple disease recording and stratification by infection intensity. Inf. Rep.BC-R-8. Canadian Forest Service, Pacific Forest Research Centre. 16 p.Bloomberg, W.J.; Cumberbirch, P.M.; Wallis, G.W. 1980. A ground survey method forestimating loss caused by Phellinus weirli, I. Development of survey design. Inf. Rep. BCR-3. Canadian Forest Service, Pacific Forest Research Centre. 24 p.Bloomberg, W.J.; Morrison, D.J. 1989. Relationship of growth reduction in Douglas-fir toinfection by Armillaria root disease in southeastern British Columbia. Phytopathology. 79:482-487.Byler, J.W.; Marsden, M.A.; Hagle, S.K. 1990. The probability of root disease on the LobNational Forest, Montana. Canadian Journal of Forest Research. 20: 987-994.Byler, J.W.; Marsden M.A.; Hagle, S.K. In: Cooley, S., ed. Western international forestdisease conference: Proceedings; 1986 September 8-12; Juneau, AK. Portland, OR: U.S.Department of Agriculture, Forest Service, Pacific Northwest Region, State and PrivateForestry. 1987: 52-56.Cobb, F.W. Jr. 1989. Interactions among root disease pathogens and bark beetles inconiferous forests. In: Morrison, D.J., ed. Proceedings of the 7th international conferenceon root and butt rots; 1988 August 9-16; Vernon and Victoria, B.C. Victoria, B.C.:International Union of Forestry Research Organizations: 142-148.Dubreuil, S.K. 1981. Occurrence, symptoms, and interactions of Phaeolus schweinitzii andassociated fungi causing decay and mortality of conifers. Ph.D. dissertation, University ofIdaho, Moscow, ID.Filip, G.M. 1977. An Armillaria epiphytotic on the Winema National Forest, Oregon. PlantDisease Reporter. 61: 708-711.Filip, G.M. 1989. Interactions among root diseases and agents of defoliation. In: Morrison,D.J., ed. Proceedings of the 7th international conference on root and butt rots; 1988 August9-16; Vernon and Victoria, B.C. Victoria, B.C.: International Union of Forestry ResearchOrganizations: 149-155.Fiip, G.M.; Goheen, D.J. 1982. Tree mortality caused by root pathogen complex inDeschutes National Forest, Oregon. Plant Disease. 66: 240-243.Filip, G.M.; Goheen, D.J. 1984. Root diseases cause severe mortality in white and grand firstands of the Pacific Northwest. Forest Science. 30: 138-142.Garrett, S.D. 1960 Rhizomorph behaviour in Armillaria mellea (Fr.) Quel., III. Saprophyticcolonization of woody substrates in soil. Annals of Botany. 24: 275-285.116Garrett, S.D. 1970. Pathogenic Root-Infecting Fungi. Cambridge: Cambridge UniversityPress. 294 p.Gregory, S.C. 1985. The use of potato tubers in pathogenicity studies of Armillaria isolates.Plant Pathology. 34: 41-48.Greig, B.J.W.; Strouts, R.G. 1983. Honey fungus. Arboricultural Leaflet 2. Revised. GreatBritain: Her Majesty’s Stationery Office; Department of the Environment, ForestryCommission. 16 p.Guillaumin, J.J.; Mohammed, C.; Berthelay, 5. 1989. Armillaria species in the norhterntemperate hemisphere. In: Morrison, D.J., ed. Proceedings of the 7th internationalconference on root and butt rots; 1988 August 9-16; Vernon and Victoria, B.C. Victoria,B.C.: International Union of Forestry Research Organizations: 27-43.Hadfield, J.S. 1984. Root disease problems and opportunities in the interior Douglas-fir andgrand fir forest types. In: Proceedings of a symposium on silvicultural managementstrategies for pests of the interior Douglas fir and grand fir forest types. Spokane, WA:University of Washington Publications: 59-66.Hagle, S.K. 1985. Monitoring root disease mortality. Report No. 85-27. Missoula, MT: U.S.Department of Agriculture, Forest Service, Northern Region, State and Private Forestry. 13p.Hagle, S.K.; Goheen, D.J. 1988. Root disease response to stand culture. In: Proceedings ofthe future forests of the intermountain west: a stand culture symposium. Gen. Tech. Rep.INT-243. Ogden, UT: U.S. Department of Agriculture, Forest Service, IntermountainForest and Range Experiment Station: 303-309.Hansen, E.M.; Goheen, D.J. 1989. Root disease complexes in the Pacific Northwest. In:Morrison, D.J., ed. Proceedings of the 7th international conference on root and butt rots.1988 August 9-16; Vernon and Victoria B.C. Victoria BC: International Union of ForestryResearch Organizations: 129-141.Hintikka, V. 1974. Notes on the ecology of Armillariella mellea in Finland. Karstenia 14: 12-31.Hobbs, S.D.; Partridge, A.D. 1979. Wood decays, root rots, and stand composition along anelevation gradient. Forest Science. 25: 3 1-42.Hood, l.A.; Morrison D.J. 1984. Incompatibility testing of Arinillaria isolates in a woodsubstrate. Canadian Forestry Service Research Notes. 4:8-9.Hood, l.A.; Redfern, D.B.; Kile, G.A. Armillaria in planted hosts. In: Shaw, C.G., III; Kile,G.A. 1991. Armillaria root disease. Agriculture Handbook No. 691. U.S. Department ofAgriculture, Forest Service. 233 p.James, R.L.; Stewart, C.A.; Williams, R.E. 1984. Estimating root disease losses in thenorthern Rocky Mountain national forests. Canadian Journal of Forest Research. 14: 652-655.Ketcheson, M.V.; Braumandi, T.F.; Meidinger, D. ;[and others]. 1991. Interior cedar -hemlock zone. In: Meidinger, D.; Pojar, 3. eds. Ecosystems of British Columbia, SpecialRep. Ser. No. 6. British Columbia: Ministry of Forests, Research Branch. 330 p.117Kile, G.A. 1980. Behaviour of an Armillaria in some Eucalyptus obliqua - Eucalyptus regnansforests in Tasmania and its role in their decline. European Journal of Forest Pathology. 10:278-296.Kile, G.A.; McDonald, G.I.; Byler, J.W. 1991. Ecology and disease in natural forests. In:Shaw, C.G.,ffl; Kile, G.A. 1991. Armillaria Root Disease. Agriculture Handbook No.691. U.S. Department of Agriculture, Forest Service. 233 p.Korhonen, K. 1978. Interfertility and clonal size in the Armillaria mellea complex. In: Shaw,C.G., Ill; Kile, G.A. 1991. Armillaria root disease. Agriculture Handbook No. 691. U.S.Department of Agriculture, Forest Service. 233 p.Kuthavy, D.L.; Partridge, A.D.; Stark, R.W. 1984. Root diseases and blister rust associatedwith bark beetles (Coleoptera: Scollytidae) in western white pine in Idaho. EnvironmentalEntomology. 13: 813-817.Lloyd, D.A.; Angove, K.; Hope, G.; Thompson, C. 1990. A guide for site identification andinterpretation of the Kamloops Forest Region. 2 vol. British Columbia: Ministry of Forests,Land Management Handbook. No. 23. Victoria, British Columbia.McDonald, G.I.; Martin, N.E.; Harvey, A.E. 1987. Armillaria in the Northern Rockies:pathogenicity and host susceptibility on pristine and disturbed sites. Res. Note INT-371.Ogden UT: U.S. Department of Agriculture, Forest Service Intermountain ResearchStation. 5p.McDonald, G.I. 1990. Relationships among site quality, stand structure, and Armillaria rootrot in Douglas-fir forests. In: Interior Douglas-fir: the species and its management:Proceedings of the symposium. Pullman: Washington State University, CooperativeExtension. lip.Meidinger, D.; Pojar, J. 1991. Ecosystems of British Columbia. British Columbia Ministry ofForests. Special Report Series, ISSN 0843-6452; no. 6 330 p.Morrison, D.J. 1972. Studies on the biology of Armillaria mellea. Cambridge: University ofCambridge. 169 p. Ph.D. dissertation. [Internal Report BC-30. Victoria, BC: CanadianForestry Service, Pacific Forest Research Centre. 169 p.].Morrison, D.J. 1981. Armillaria root disease. A guide to disease diagnosis, development andmanagement in British Columbia. Information Report BC-X-203. Environment Canada,Canadian Forestry Service: 1-16.Morrison, D.J. 1989. Pathogenicity of Armillaria species is related to rhizomorph growthhabit. In: Morrison, D.J., ed. Proceedings of the 7th international conference on root andbutt rots; 1988 August 9-16; Vernon and Victoria, BC. Victoria, BC: International Unionfor Forestry Research Organizations: 584-589.Morrison, D.J.; Chu, D.; Johnson, A.L.S. 1985. Species of Armillaria in British Columbia.Canadian Journal of Plant Pathology. 7: 242-246.Morrison, D.J.; Merler, H.; Norris, D. i991a. Detection, recognition and management ofArmillaria and Phellinus root diseases in the southern interior of British Columbia. FRDAII Rep. British Columbia: Ministry of Forests, Research Branch. 25 p.Morrison, D.J.; Wallis, G.W.; Weir, L.C. 1988. Control of Armillaria and Phellinus rootdisease: 20-year results from the Skimikin stump removal experiment. Information ReportBC-X-302. Canadian Forestry Service, Pacific Forestry Centre. 16 p.118Morrison, D.J.; Williams, R.E.; Whitney, R.D. 1991b. Infection, disease development,diagnosis and detection. In: Shaw, C.G., III; Kile, G.A. 1991. Armillaria root disease.Agriculture Handbook No. 691. U.S. Department of Agriculture, Forest Service. 233 p.Morrison, D.J.; Pellow, K. 1993. Development of Armillaria root disease in a 25-year-oldDouglas-fir plantation. In: Proceedings of the 8th international conference on root and buttrots; Aug. 9-16, 1993. W.K. Sweden and Haikko, Sweden. International Union of ForestryResearch Organizations: 560-571.Reaves, J.L.; Shaw, C.G., III; Mayfield, J.E. 1990. The effects of Trichoder,na spp. isolatedfrom burned and non-burned forest soils on the growth and development of Armillariaostoyae in culture. Northwest Science. 64: 39-44.Redfern, D.B.; Filip, G.M. 1991. Inoculum and infection. In: Shaw, C.G., III; Kile, G.A.1991. Armillaria root disease. Agriculture Handbook No. 691. U.S. Department ofAgriculture, Forest Service. 233 p.Rishbeth, J. 1972. The production of rhizomorphs by Armillaria mellea from stumps.European Journal of Forest Pathology. 2: 193-205.Rishbeth, J. 1985. Arinillaria: resources and hosts. In: Moore, D.; Casselton, L.A.; Wood,D.A.; [and others], eds. Developmental biology of higher fungi. Cambridge: CambridgeUniversity Press: 87-101.Shaw, C.G., III; Roth, L. F.; Rolph, L.; [and others]. 1976. Dynamics of pine and pathogenas they relate to damage in a forest attacked by Armillaria. Plant Disease Reporter. 60:214-218.Shaw, C.G., III. 1975. Epidemiological insights into Armillaria mellea root rot in a managedponderosa pine forest. Corvallis, OR: Oregon State University. 201 p. Ph.D. dissertation.Stewart, C.A.; James. R.L.; Bousfield, W.E. 1982. A multi-stage sampling technique toassess root disease impact on the Clearwater and Nez Perce National Forests, Idaho.Cooperative Forestry and Pest Management Rep. 82-14. U.S. Department of Agriculture,Forest Service, Northern Region. 33 p.Taylor, S.P., ed. 1986. Forest insect and disease impacts in timber supply areas. PestManagement Report NO. 6. British Columbia: Ministry of Forests, Forest ProtectionBranch. 254 p.Thomas, H.E. 1934. Studies on Armillaria mellea (Vahl) Quel., infection parasitism and hostresistance. Journal of Agricultural Research. 48: 187-2 18.Thrower, J.S.; Goudie, J.W. 1992. Development of height-age and site index functions foreven aged, interior Douglas-fir in British Columbia. B.C. Ministry of Forests, ResearchBranch, Research Note No. 109. 22 p.van der Kamp, B.J. 1992. Rate of spread of Armillaria ostoyae in the central interior of BritishColumbia. Canadian Journal of Forest Research. 23: 1239-1241.Wallis, G.W.; Bboomberg, W.J. 1981. Estimating the total extent of Phellinus weirli root rotcenter using above- and below- ground disease indicators. Canadian Journal of ForestResearch. 11: 827-830.Wargo, P.M.; Shaw, C.G., III. 1985. Armillaria root rot: the puzzle is being solved. PlantDisease 69: 826-832.119Watling, R.; Kile, G.A.; Burdsall, H.H, Jr. 1991. Nomenclature, taxonomy, andidentification. In: Shaw, C.G., III; Kile, G.A. 1991. Armillaria root disease. AgricultureHandbook No. 691. U.S. Department of Agriculture, Forest Service. 233 p.William, R.E.; Leaphart, C.D. 1978. A system using aerial photography to estimate area ofroot disease centers in forests. Canadian Journal of Forest Research. 8: 2 14-219.Williams, R.E.; Shaw, C.G., ifi; Wargo, P.M.; [and others]. 1989. Armillaria root disease.Forest Insect and Disease Leaflet 78 (rev.). U.S. Department of Agriculture, ForestService. 8 p.Williams, R.E.; Marsden, M.A. 1982. Modelling the probability of root disease centeroccurrence in northern Idaho forests. Canadian Journal of Forest Research. 12: 876-882.Woeste, U. 1956. Anatomische Untersuchungen uber die Infektionswege einiger Wurzelpilze.[Anatomical studies on the methods of infection of several root fungi.] In: Shaw, C.G., III;Kile, G.A. 1991. Armillaria root disease. Agriculture Handbook No. 691. U.S.Department of Agriculture, Forest Service. 233 p.Zeller, S.M. 1926. Observations on infections of apple and prune roots by Armillaria melleaVahi. Phytopathology. 16: 479-484.APPENDIX121APPENIMX 1Appendix 1.1 List of Abbreviations for Tree SpeciesAbbreviation Common Name Scientific NameAt trembling aspen Populus tremuloides Michx.Bi subalpine-fir Abies lasiocarpa (Hook.) Nutt.Ep paper birch Betula papyrifera Marsh.Cw western redcedar Thuja plicata Donn ex D. DonFd Douglas-fir Pseudotsuga menziesii (Mirb.) FrancoHw western hemlock Tsuga heterophyla (Rafn.) Sarg.Lw western larch Larix occidentalis Nutt.P1 lodgepole pine Pinus contorta Dougl. ex Loud.Pw western white pine Pinus monticola Dougl. ex D. DonSe Engelmann spruce Picea engelmanni Parry ex Engelm.Appendix 1.2 Biogeodlimatic Ecological Classification System AbbreviationsAbbreviation ClassificationBEC Biogeoclimatic Ecological ClassificationICR Interior Cedar Hemlock zoneIDF Interior Douglas-fir zoneESSF Engelmann Spruce Subalpine fir zoneMS Montane Spruce zoneSBS Sub-boreal Spruce zoneIDFmwl Shuswap moist warm Interior Douglas-fir variantIDFmw2 Thompson moist warm Interior Douglas-fir variantICHmw2 Shuswap moist warm Interior Cedar Hemlock variantICHmkl Kootenay moist cool Interior Cedar Hemlock variantAppendix 1.3 List of Acronyms Used for Data AnalysesAbbreviation DefinitionDRC diameter at root collarARMSEVB proportion of stumps infected prior to harvestARMSEVA proportion of stumps infected both before and after harvestSTMPBEFR number of stumps infected before harvestSTMPAFTR total number of stumps infectedBABEFOR total basal area in m2/ha of stumps pre-harvestBAAFTER total basal area in m2/ha of stumps infectedCONSTEMS number of healthy conifer regeneration/haPROINFCT proportion of conifer regeneration infected out of total coniferregenerationTREINFCT number of conifer stems infectedInfected refers to infections with Armillaria ostoyae