FOREST SCIENCE PROGRAM. Reducing the impact of Armillaria root disease. via mixed species plantations including western red cedar

Transcription

1 FOREST SCIENCE PROGRAM Reducing the impact of Armillaria root disease via mixed species plantations including western red cedar Annual Technical Report FSP Research project Y M. R. Cleary and B.J. van der Kamp Department of Forest Sciences, UBC April 22, 2005

2 2 ABSTRACT: Necrophylactic periderm formation and compartmentalization of infected tissue was studied in tissue samples collected from the roots of year old western red cedar (Thuja plicata), western hemlock (Tsuga heterophylla)and Douglas-fir (Pseudotsuga menziesii) trees infected with Armillaria ostoyae. Following penetration by the fungus, a higher frequency of resistance reactions were induced in western red cedar trees than in to western hemlock and Douglas-fir. Resistance reactions in cedar involved necrophylactic periderm formation around a site of penetration. Subsequently, a hypersensitive response was induced involving rhytidome formation which extended for some distance proximally and distally from a recently formed periderm. Effective compartmentalization following cambial invasion also limited the extent of girdling by the fungus. In a survey of juvenile mixed species plantations throughout the southern interior of B.C., only 2% of the western red cedar trees were killed by A. ostoyae compared to 25% of the Douglas-fir trees. The frequency of western red cedar trees showing effective compartmentalization at the root collar was significantly higher than in Douglas-fir trees. Although the risk of mortality decreased with increasing tree size for both western red cedar and Douglas-fir trees, the rate of decrease was noticeably greater in cedar than in Douglas-fir. Results indicate that western red cedar is more resistant to A. ostoyae than other conifers and that the inclusion of cedar in higher proportions when planting infested sites may reduce the overall impact of Armillaria root disease.

3 3 INTRODUCTION: Armillaria root disease is a significant health concern in the southern interior forests of British Columbia. Armillaria ostoyae (Romagn.) Herink, the fungus causing the disease, is most predominant in the Interior Cedar-Hemlock (ICH) biogeoclimatic zone where the incidence of diseased trees in mature stands can be as high as 80% (Morrison et al. 2001). After harvest, the fungus will colonize stump and root systems and become inoculum. Mortality in the new regeneration starts at around age 5-7 years when the roots of planted trees contact residual inoculum. Cumulative mortality can be as much as 20% by age 20 years resulting in undesirable stocking in juvenile stands (Morrison and Pellow 1994). Previous studies also show that within the ICH zone, above-ground symptoms of Armillaria can be detected in only one-quarter of the trees with belowground infection (Morrison et al. 2000). Actual incidence of diseased trees can be as much as 35% at age 20 and growth loss will occur in trees sustaining non-lethal infections (Morrison et al. 2000). Ultimately, these losses can become considerable over time and create serious challenges for managing sustainable timber production on infested sites. Few options are available to mitigate potential losses due to Armillaria root disease. Removal of stumps from the ground is a very effective means of reducing the amount of woody inoculum that would otherwise be available to the fungus. Another less intrusive option is to plant conifer species that have a low susceptibility to killing by A. ostoyae. Data obtained from inoculation trials initiated in 2002 reported that the frequency at which resistant reactions (including necrophylactic periderm formation and compartmentalization) are induced following invasion by the fungus is significantly higher in western red cedar (Thuja plicata Donn ex D. Don) trees than in Douglas-fir (Pseudotsuga menziesii (Mrib) Franco) and western hemlock (Tsuga heterophylla (Raf.) Sarg.) trees (Cleary et al. 2004). Indirect evidence from previous studies (DeLong 1996, Morrison et al. 2000) and preliminary evidence under the current study suggests western red cedar is more resistant to infection by A. ostoyae than other common conifers. When roots become infected by A. ostoyae, defense mechanisms on root systems will determine whether the reaction is one of susceptibility or resistance. Several studies have shown necrophylactic periderm to be involved in resistant reactions in woody plants in response to mechanical injury or pathogenic invasion (Blanchette and Biggs 1992; Wahlstrom and Johansson 1992; Robinson and Morrison 2001). Some of this resistance is attributed to structural characteristics in the necrophylactic periderm which may inhibit further spread of A. ostoyae in adjacent tissue, confining it to a lesion in the bark (Robinson and Morrison 2001). The methodology developed by Robinson (1996) to describe host response to infection by A. ostoyae on western larch and Douglas-fir will be used in our investigation to describe host response to infection in western red cedar. Our work to date has shown some resistant mechanisms operating in western red cedar that are effective at containing infections and halting the spread of the fungus to adjacent healthy tissue. However, before recommending the use of western red cedar as a significant stand component when regenerating Armillaria-infested sites, it was necessary to show that the greater resistance observed in western red cedar at the individual root contact level translates into reduced infection and mortality at the stand level. In 2004, we surveyed juvenile mixed

4 conifer stands throughout the ICH in the southern interior of B.C. to examine symptom development and mortality rates caused by A. ostoyae in western red cedar and compare that to a known susceptible species, Douglas-fir. The aim of this research project was two-fold: First, to complete a study of natural resistance mechanisms in the roots of western red cedar to demonstrate that cedar does indeed exhibit an effective response against Armillaria. Secondly, to demonstrate through a survey of juvenile mixed conifer stands that the presence of these natural resistance mechanisms leads to mortality rates in cedar that are significantly lower than in other common conifers. This study will yield information about natural resistant mechanisms in western red cedar and concomitantly provide information on the relative susceptibility of other conifers including western hemlock and Douglas-fir. Results will also determine whether the inclusion of western red cedar among susceptible conifers like Douglas-fir, in natural or artificially regenerated clearcuts will reduce the overall impact of Armillaria root disease. 4

5 5 1. Study of natural resistance mechanisms METHODS: Site selection for inoculation trials In 2004, two field inoculation trials were implemented to examine host response to infection by A. ostoyae on the roots of western red cedar, western hemlock and Douglas-fir. Study sites were selected in the Kingfisher area, east of Enderby in the Vernon Forest District and near Nakusp in the Arrow Lakes/Boundary Forest District. Both sites were in the moist-warm subzone of the ICH biogeoclimatic zone in the southern interior of B.C. Inoculum block preparation and technique: Inoculum consisted of fully colonized segments of garry oak (Quercus garryana Dougl.) branches that were prepared several months previously in the lab. A. ostoyae isolate (provided by D. Morrison, CFS) was used to inoculate garry oak segments. In the field, tree roots were carefully exposed so as to not injure the outer bark tissue and trigger a host response independent of Armillaria infection. At least 10 individual trees of Douglas-fir, hemlock and cedar were inoculated. A single root was inoculated with one inoculum block and when possible, all inoculations on a single tree (up to 4) occur on individual roots. Existing rhizomorphs were removed from inoculum blocks in order to stimulate new rhizomorph growth and a block was placed either directly below or alongside the root without injuring the root itself and then covered with mineral soil. Uninfected segments of garry oak were used as controls. Control blocks were placed on the same tree but different lateral roots. Trees were tagged and the following information was recorded: tree species, DBH, root diameter, and distance of the inoculum block from the root collar. Tree roots respond to mechanical wounds (or any factor that kills or removes bark) by the formation of barriers in bark and wood in ways that are parallel to those observed in response to Armillaria invasion. These responses are similar in all species. Hence in order to understand differences in response to Armillaria invasion, it was useful to have mechanical wounds as a control, and to look for phenomena (responses) specifically related to the presence of Armillaria. Abiotic wounding on healthy roots was performed without physically disrupting the outer bark. Shurfreeze cryospray was applied to the exterior surface of the root for a given length of time depending on the thickness of the bark. This method achieved a freezing injury in the inner bark tissue of various degrees and thereby permitted characterization of phellogen restoration and host response in the absence of Armillaria. Field methods and sampling technique for freezing injuries were similar to that of inoculations and controls. Each treatment (inoculation, freezing and control) had at least 15 replications for each species. Field sampling All trees were inoculated and abiotically wounded in early May A small sample of treated roots from each species were harvested at 9- and 11- weeks from the Nakusp and Kingfisher sites, respectively. A larger sample of roots were harvested from both sites at 5- months after inoculation. At each sampling date, root systems were carefully excavated and checked for infection at inoculum contact and any rhizomorph-initiated lesions occurring further along the length of the root. Macroscopic observations of host response (presence of a

6 necrophylactic periderm, resin exudation, compartmentalization and callusing), lesion size, root age, tree age and the current radial increment were recorded at the time of sampling. Wood samples showing compartmentalization and callusing were sectioned from the root and preserved in FAA. Bark samples from both Armillaria-lesions and freezing injuries were dissected to reveal the radial face of the infection front. An infection front was recognized by the presence of either (1) a necrophylactic periderm separating healthy (uncolonized) inner bark from infected (colonized) inner bark, or (2) browning of tissue in advance of penetrating mycelium. Additional samples were collected from the control treatment at the point of contact with the block. Bark samples were coated with Optimum Cutting Temperature (OCT) compound and immediately stored in liquid nitrogen. The cryofixation of tissue samples allows for direct observation of cell contents in their natural state at some later date using fluorescence and other modes of illumination. Microscopic examination of host response In the lab, bark samples were dissected in a cryostat set at 20 o C. Frozen sections mounted on slides were examined microscopically under tungsten illuminated bright field (BF) and a mercury lamp with fluorescence filter combinations on a freezing stage set at -35 o C. Microscopic observation of infected bark tissue in this context permits physiological interpretation of any anatomical changes of cells associated with infected tissue based on changes in fluorescence characteristics when compared with healthy tissue. Additional sections were obtained, thawed and then stained for lignin (using phloroglucinol + HCl) and suberin (using Sudan III or visualization by autofluorescence after staining with phloroglucinol + HCl). Photomicrographs were taken of frozen and stained cyrostat sections. Wood sections were embedded in paraffin and sectioned on a rotary microtome to describe barrier zone anatomy associated with compartmentalization of infected tissue. RESULTS Examination of roots at 9- and 11-weeks, from the Nakusp and Kingfisher trials, respectively, revealed a low proportion of inoculated roots showing successful penetration of the outer bark tissue for all species. It was determined that infection was still in the early stages and at least another 2 months were required in order to see any expression of resistance in host species in response to the invading fungus. Roots were subsequently harvested at five months after inoculation at which time over 80% infection occurred in both the western red cedar and western hemlock roots and over 50% in the Douglas-fir roots. Host response in Douglas-fir and Western hemlock The mode of penetration on Douglas-fir roots appears to involve the suppression of phellogen activity underlying a penetrating rhizomorph. Bark samples showing only rhizomorphs adhering to the outer surface of the root, but lacking penetration of the inner bark, showed a narrow zone of phellogen activity (typically 1-cell wide) compared to an more active phellogen zone (2-4 cells wide) on either side of an overlying rhizomorph (Figure 1.1). This type of host-recognition at surface contact was observed more frequently in Douglas-fir than in the other conifers studied. As the fungus degrades the outer suberized rhytidome, a lateral branch will develop and penetrate the host tissue as a larger unit rather than as individual hyphae (Figure 1.2) 6

7 7 Rh Rh Figure 1.1. A radial section of a Douglas-fir root with a rhizomorph attached to the outer surface of the root. The living bark underlying the rhizomorph showed a significant reduction in phellogen activity (arrows); viewed with Blue-Light (BL) excitation; Rh, rhizomorph. Figure 1.2 A radial section of a western hemlock root showing penetration of the living bark by a rhizomorph. Necrosis occurs ahead of mycelial penetration. viewed with bright field (BF) Douglas-fir and western hemlock roots infected by A. ostoyae typically showed susceptible host reactions consisting of no visible host response at the infection front. Adjacent phloem often appeared brown and either lacked significant hypertrophy or appeared invariably hypertrophied. At times, sporadic lignification of adjacent phloem tissue was seen in advance of a penetrating mycelium which indicated the initiation of non-suberized impervious tissue (NIT), a prerequisite for necrophylactic periderm (NP) formation. However, distinct zones of NIT were frequently lacking and/or the fungus penetrated beyond the developing NIT zone. The percentage of infections on western hemlock and Douglas-fir roots with no NP formation and/or compartmentalization to contain the fungus was significantly lower compared to western red cedar. BPh Rh APh Figure 2.1. A macrophotograph of a Douglas-fir root showing large wedges of mycelium colonizing the inner bark tissue. Browning of phloem can be seen around the penetrating mycelium; Rh, rhizomorph Figure 2.2. A radial section showing the lack of significant hypertrophy at the infection front. The fungus was able to preclude the formation of a necrophylactic periderm to contain the infection; BL; BPh, brown phloem; APh, adjacent phloem.

8 At times, a necrophylactic periderm was formed to contain an infection but then breached by the fungus (Figure 3). Cryofixation revealed sclereids as a possible area of weakness in the phloem (Figure 4). Few samples showed compartmentalization following cambial invasion by the fungus. The uninjured cambium adjacent to killed tissue formed traumatic resin canals.. NP 8 3 Figure 3. A macrophotograph of a western hemlock root. An initial attempt at containing the infection within a necrophylactic periderm (NP) was unsuccessful. The fungus breached this barrier resulting in necrosis down to the vascular cambium (arrow). Scl NP Scl 4.1 TRD 4.2 Figure 4.1. A radial section of a western hemlock root showing incomplete necrophylactic periderm (NP) formation around clusters of sclereids (Scl); BL. Figure 4.2 A radial section of a western hemlock root showing traumatic resin ducts (TRD) in the wood adjacent to cambial invasion; BL. The frequency at which susceptible reactions occurred in both Douglas-fir and western hemlock roots inoculated with A. ostoyae is consistent with previous field inoculation trials documenting the lack of successful resistance responses in both species.

9 Host response in Western red cedar Results from inoculation trials in 2004 were consistent with earlier field trials documenting a higher percentage of resistance reactions following invasion by A. ostoyae. The percentage of cedar root inoculations resulting in successful penetration by the fungus was similar to Douglas-fir and western hemlock, however host reactions following invasion by the fungus were markedly different. There was no difference in the frequency of infections and host reactions for each species between sites. Cedar frequently responded to infection by Armillaria by forming a necrophylactic periderm deeper in the bark tissue. This barrier appeared to be effective at preventing the spread of A. ostoyae to adjacent healthy phloem (Figure 5). Microscopic examination of necrophylactic periderm formation revealed phellem cells wrapped around cedar phloem fibers (Figure 6.1 and 6.2).. 9 NP 5 Figure 5. A macrophotograph showing a necrophylactic periderm (NP) surrounding necrotic tissue caused by invasion by A. ostoyae on a western red cedar root. BPh BPh APh APh Figure 6.1. A radial section of a western red cedar root showing complete necrophylactic periderm (arrow) formation separating necrotic from healthy phloem tissue; BF; BPh, APh. Figure 6.2 The same radial section viewed under fluorescence showing the phellem of the newly formed periderm wrapped around phloem fibers; BL.

10 Further examination of lesions revealed a unique phenomenon occuring in western red cedar roots which involves a type of localized hypersensitive response in the vicinity of a site of wounding. This response occurs after the initiation of phellogen renewal and necrophylactic periderm formation and appears to be non-specific, occurring in both abiotically wounded and inoculated roots (Figure 7) st NP Figure 7. A macrophotograph 7 of a western red cedar root showing a localized response (arrow) induced after the formation of a necrophylactic periderm. Adjacent to an area of injured tissue, a large zone of phloem parenchyma cells become hypertrophied and undergo changes in fluorescence characteristics which are indicative of the early stages of dedifferentiation (Figure 8). This response is consistent with normal rhytidome formation and typically occurs on either side of the lesion extending for some distance beyond the primary wounded tissue (Figure 9). The intensity of the localized response that is expressed (e.g. the lesion size or the length at which rhytidome formation occurs beyond the obvious infection front), depends on the extent of colonization of the bark tissue by the fungus. Hence, when cedar bark is injured, the host will form a necrophylactic periderm to confine the infection and then stimulate rhytidome production on either side of the lesion so that eventually one continuous rhytidome layer will be formed which may eventually be sloughed from the surface of the root. Zone of hypertrophy 8 phellem Figure 8. A radial section of a western red cedar root showing extensive hypertrophy and dedifferentiation in the adjacent phloem tissue near the primary infected tissue. New phellem cells are differentiating immediately abutting zone of hypertrophy. BL.

11 11 9 Rhytidome Figure 9. A macrophotograph showing rhytidome formation in cedar as a large zone of hypertrophied phloem with a newly differentiating layer of phellem (pigmented line) in the inner bark. Lesions to the cambium were frequently compartmentalized (Figure 10). Immediately adjacent to an area of killed cambium, the uninjured cambium produced a higher than average number of axial parenchyma that accumulated dense phenolic deposits. This host reaction was very effective at stopping the spread of the fungus and limiting the extent of cambial invasion and/or girdling in roots. Rh NP 10 Figure 10. A macrophotograph showing necrosis down to the vascular cambium (arrow). Compartmentalization of infected tissue occurred and lateral in-growth of callus was evident from either side of the killed cambium. Note rhizomorph (Rh) embedded within the necrotic zone, also bound by a necrophylactic periderm (NP). There were noticeable differences in the frequency of successful resistance reactions in response to invasion by A. ostoyae between species. Results indicate that western hemlock and Douglas-fir roots are more susceptible to infection by A. ostoyae than western red cedar. Observations from field inoculation trials suggest that the nature of resistance in western red cedar against infection by A. ostoyae includes both effective resistance responses formed immediately abutting an advancing infection front (i.e. necrophylactic periderm formation and the accumulation of phenolics in axial parenchyma during the process of compartmentalization), as well as the long-range resistance response involving rhytidome formation.

12 2. Symptom development and mortality rates by species in juvenile stands METHODS: Site selection Participation of all project team members and industry partners yielded lists of potential candidate sites for surveying juvenile mixed conifer stands. Candidate sites were visited and selected based on the following criteria 1. Stands planted to Douglas-fir or some mixture of Douglas-fir and other conifers between 1979 and 1989 with a uniform and intimate mixture of Douglas-fir and cedar 2. Significant ( 30%) western red cedar component 3. Stocking between 2,000-10,000 trees per hectare 4. No juvenile spacing 5. Accessible by main logging road 6. Moderately to heavily infected with A. ostoyae Plot selection and measurement of trees At each site a transect line was established in a random direction at least 20 m from the stand edge. From the starting point, any tree killed by A. ostoyae that was encountered within 3 m of the transect line was considered a potential plot centre tree. A 10 m radius plot was centred around a potential plot centre tree provided a second set of criteria (listed below) were satisfied within the plot. a. At least 10% infection by A. ostoyae (lethal and non-lethal) in conifers b. Stocking of western red cedar comprises approximately 30% of the species composition c. Number of trees tallied would yield between trees per plot Infected trees along the transect line were considered for centre tree until a suitable plot location was found. Within a plot, each tree (living and dead) and 3 cm DBH was identified and recorded by species, tree size (DBH), crown class, and disease status. Armillaria infection was confirmed by sub-cortical mycelial fans or impressions of such fans in resin-soaked and/or necrotic bark in all trees suspected of being infected or killed by A. ostoyae. The plot center tree was not part of the sample. The root collar was examined on all trees for evidence of old or current basal lesions. Non-lethal infections at the root collar were classified as either progressive (eg. the fungus was advancing in the inner bark and cambial tissue) or callused (the fungus was compartmentalized and the spread of the fungus had been stopped). The search for the next plot started 50 m along the transect line from where the last plot was located to ensure no overlap of adjacent plots. The number of plots obtained for each site largely depended on the number of acceptable plots that fit within the criteria listed above, the number of trees per plot and the length of time required to survey each plot. Statistical analysis The proportion of trees killed or infected by Armillaria was determined by species for each plot and site. Statistical analysis was done using SPSS statistical package (SPSS Inc., Chicago, IL). Frequency data for species mortality was analyzed by Chi square tests. A logistic regression model was fitted to the data [ie. log p/(1-p) = a + b (dbh)] whereby p is the probability of mortality and a and b regression co-efficients depend on both species. The Hosmer- Lemeshow goodness of fit test was used to test whether the model fit the data adequately. 12

13 13 RESULTS: Experimental sites were selected from among candidates throughout the ICH in the Northern Okanagan/Shuswap and Arrow-Boundary Districts (Figure 11). Source: Figure 11. Location of study sites (1-20) used for surveying species mortality caused by Armillaria root disease in juvenile mixed conifer stands. Table 1 lists the characteristics of all sites selected for surveying. The sample population is comprised of 51 plots across 20 sites and a total of 7175 trees. Mean proportions for Douglasfir and western red cedar were relatively similar at 33% and 31.3% respectively, however, tree size distribution differed between species (Figure 12).

14 14 Table 1. Characteristics of the twenty sites surveyed for disease incidence in the North Okanagan/Shuswap and Arrow/Boundary Forest Districts in the southern interior of B.C. Site Name Mapsheet Stand Polygon BEC Stand Species No. of Total No. % Fd % Cw % Hw % Above- % Cumulative No. Opening No. Zone/ subzone Age Composition Plots of trees Ground Infection Mortality 1 Arrow Park 82K /S ICHmw2 15 PwFC(H) Baird Lake 82L ICHmw2 18 FE(LC) Erie Creek 82F /S ICHmw2 16 FCAc(Pw) Hidden Lake 82L ICHmw2 28 EFC(HAc) Humamilt Lake 82M /S ICHmw3 21 AcCF(H) Kidney Lake 82L ICHmw2 22 FCH(E) Kuskanax 82K /S ICHmw2 24 SFPwAc(C) Mabel Lake_Icebox 82L ICHmw2 20 FCH(L) Mabel Lake_Simard 82L /S ICHmw2 21 EF(CH) Mabel Lake_2 82L ICHmw2 16 EFC * Monashee Creek 82L ICH mk1 21 CF(H) Northfork_55 82M /S ICHmw3 17 CHF Northfork_60 82M /S ICHmw3 21 CFH Toledo 82L /S ICHmw2 23 CH(FPwAcE) Trinity Valley 82L ICHmw2 15 FEAcC Trout Lake_1 82K /S ICHmw2 21 CFPw(H) Trout_Lake_2 82K /S ICHmw2 25 FC Trout Lake_3 82K /S ICH wk1 25 FC(H) Wilson Creek 82K /S ICHmw2 17 EF(C ) Worthington 82E_ /S ICHmw2 33 HSFC TOTAL: * = naturally regenerated stand

15 15 Western red cedar Douglas-fir Figure 12. Size distribution as number of trees per DBH class of western red cedar and Douglas-fir trees for all 20 sites combined.

16 16 Table 2. Total number of conifer trees tallied by species in different disease status categories CONIFER SPECIES a DISEASE STATUS Fd Hw Cw Bg Lw Pl Se Pw TOTAL Healthy Infected by A. ostoyae (PROGRESSIVE infection at the root collar) Infected by A. ostoyae (CALLUSED infection at the root collar) Dead (KILLED by A. ostoyae) b Dead (unknown/other factors) c 53 Total a Fd = Douglas-fir; Hw = western hemlock; Cw = western red cedar; Bg = grand fir; Lw = western larch; Pl = lodgepole pine; Se = Englelmann spruce; Pw = western white pine. b Western larch was found at just a few sites, however the percentage of western larch trees killed by A. ostoyae (23%, n=90) was comparable to Douglas-fir (25%, n=2396). c Incidence of mortality caused by Cronartium ribicola (J.C. Fisch) was 19% of the total number of western white pine trees killed. There were significant differences (p < ) in the incidence of mortality among species over all sites. Cumulative mortality ranged between 4.3% and 32.2 % with a mean average of 11.4% (Table 1). However, this survey was not designed to yield information about overall amount of Armillaria infection, nor was it designed to capture the variation in incidence of infection by plots or site. In our search for candidate sites, we deliberately sought out moderately to heavily infected areas with the appropriate mixture of the species of interest (Douglas-fir and western red cedar) in order to make direct comparisons of symptom development and mortality rates between the two species. Table 2 shows the total number of conifer trees tallied for all sites in different disease status categories. Disease status and tree measurements were also recorded for hardwood species at each site (Appendix 1).

17 17 The frequency of mortality for a particular host species is a direct measure of its susceptibility to killing by A. ostoyae and the ability of that species to survive in the presence of inoculum. Significant differences were found in the percentage of trees killed by A. ostoyae between species across all sites surveyed (Table 3). Table 3. Proportion of the total number of trees by disease status category for Douglas-fir, western hemlock and western red cedar. SPECIES DISEASE STATUS Fd Hw Cw Proportion of trees Healthy Infected by A. ostoyae (PROGRESSIVE infection at the root collar) Infected by A. ostoyae (CALLUSED infection at the root collar) Dead (KILLED by A. ostoyae) Dead (unknown/other factors) tr a tr tr Total (n=2396) (n=1220) (n=2169) = proportion of trees killed by unknown/other factors was less than 0.5% a tr Twenty-five percent (n=2396) of the Douglas-fir trees were killed by Armillaria compared to only 2% (n=2169) of the western red cedar trees (p<0.0001). The logistic model which compared the probability of mortality between Douglas-fir and western red cedar showed species to be significant. This infers that the probability of western red cedar trees being killed by Armillaria is significant lower than Douglas-fir trees (β = , p < ). The Hosmer-Lemeshow test was not significant which suggests that the logistic model adequately fits the data (χ 2 = 3.733, p = 0.880). The model calculated the corresponding odds ratio which indicates that in young juvenile stands the odds of a Douglas-fir being killed (Exp β ) is 14.6 times greater (95% C.I. = 10.57, 20.71) than a western red cedar tree.

18 18 The percentage of infected trees with progressive lesions at the root collar differed among species (Table 4). Three percent of the infected western red cedar trees (n=117) had progressive lesions at the root collar compared to 18% (n=752) and 33% (n=268) of the Douglas-fir and western hemlock trees, respectively. Most lesions of this type are susceptible host reactions showing no visible host response at the infection front in the form of necrophylactic periderm formation. The smaller percentage of progressive lesions at the root collar in cedar compared to the other conifers suggests that fewer cedar trees are at risk to being killed by A. ostoyae. Table 4. Incidence of progressive infections, callused infections and mortality in Douglas-fir, western hemlock and western red cedar as a proportion of the total number of trees with above-ground signs or symptoms of A. ostoyae. SPECIES DISEASE STATUS Fd Hw Cw Proportion of trees Infected by A. ostoyae (PROGRESSIVE infection at the root collar) Infected by A. ostoyae (CALLUSED infection at the root collar) Dead (KILLED by A. ostoyae) Total (n=752) (n=268) (n=117) Compartmentalization and callusing at the root collar also differed among species (Table 4). Sixtyone percent (n=117) of the western red cedar trees showed effective compartmentalization of Armillaria infections compared to 3% (n=752) and 19% (n=268) of the Douglas-fir and western hemlock trees, respectively. A higher frequency of compartmentalized lesions in cedar is consistent with our earlier work looking at the response of cedar to root inoculations with A. ostoyae which showed a higher frequency of resistance reactions compared to Douglas-fir and western hemlock (Cleary et al. 2004). The ratio of incidence of mortality between western red cedar and Douglas-fir can be visualized in Figure 13 where individual data points represent sites. If both species were equally susceptible to being killed by A. ostoyae, one might expect the ratio of the incidence to fall around the green diagonal line (slope = 1). Logistic regression analysis indicated that mortality of the two species differed significantly (p < ).

19 19 Figure 13. Average mortality (percent) by site for western red cedar plotted over mortality of Douglas-fir. Results also showed that the probability of mortality among trees infected by Armillaria depends on both species and tree size (Figure 14.1). Incidence of mortality was significantly greater in the smaller diameter size class than in the larger size classes for both species and cedar mortality was consistently lower than Douglas-fir. Although the risk of mortality decreases with increasing tree size in both species, the rate of decrease is noticeably greater among cedar compared with Douglasfir trees. There was an increasing trend in the proportion of infected trees showing effective compartmentalization and callusing and tree size for both species, but the increase was markedly greater among western red cedar compared with Douglas-fir trees (Figure 14.2). In the majority of conifer plantations, cedar comes in as natural regeneration, so that trees are younger and smaller than the planted species, which could make direct comparisons of mortality difficult. It is more likely that mortality in the planted Douglas-fir has been happening for at least a few years longer than in cedar. Nonetheless, reduced mortality rates exhibited in even the intermediate size classes and the higher frequency of callusing observed in the smallest diameter size class indicates that western red cedar is more resistant to infection by A. ostoyae and this resistance occurs much earlier than most other conifers.

20 20 Fig Fd Cw n=313 MORTALITY 0.80 n=32 n= Proportion n=46 n=88 n= n=18 n= size class (DBH mm) Fig CALLUSED LESIONS n=18 n= Proportion n=46 n=32 n=36 n=88 n=313 n= Size class (DBH mm) Figure 14. Mortality (14.1) and callused lesions at the root collar (14.2) in Douglas-fir and western red cedar as a proportion of the total number of trees with above-ground signs or symptoms of A. ostoyae by size class

21 21 DISCUSSION: In this study, we investigated the host response in western red cedar trees following infection with A. ostoyae at the tissue and cellular level and compared that with other conifers including western hemlock and Douglas-fir. The resistance observed in cedar at the root contact level was assessed on a larger scale by documenting symptom development and mortality rates for cedar relative to other common conifers in juvenile mixed conifer stands throughout the southern interior region of British Columbia. Root systems of western red cedar are extensive consisting of a dense network of small diameter roots that extend for large distances in the soil. Cedar roots are usually deeper than western hemlock but shallower than Douglas-fir. Douglas-fir typically have larger welldeveloped lateral roots. Characteristics of both types of rooting habits enable both species to have a high probability of contacting potential inoculum sources in the soil because the frequency of inter-tree contacts depends on several factors including inter-tree distance, rooting depth, and most importantly, the number of roots (Reynolds and Bloomberg 1982). The overall outcome of an Armillaria infection in host species is largely determined by host resistance and the inoculum potential of the fungus. When Armillaria invades the bark tissue on roots, a host may form a necrophylactic periderm to contain the infection. Cedar appears to form these resistant reactions more frequently than other conifers. Recent examination of inoculated roots revealed a unique phenomenon occurring in cedar whereby the host initiates rhytidome formation proximally and distally to a recently formed necrophylactic periderm. This type of hypersensitive response is a non-specific reaction, occurring in both abiotically wounded and infected tissue. Successive periderm formation on either end of an Armillaria-caused lesion will extend for some distance along the root so that over time one continuous rhytidome layer will be formed which may eventually be sloughed from the surface of the root. The resulting Armillaria lesion will appear elongated, but the fungus rarely girdles the root. When the fungus penetrates to the depth of the vascular cambium, the infection is rapidly compartmentalized and contained within that barrier and it is rare that the fungus will escape this barrier once formed. Data collected in this study documenting symptom development and mortality rates for cedar is in agreement with our earlier work reporting effective root disease resistance for this species. Natural resistance mechanisms in the bark (including necrophylactic periderm formation and successive rhytidome formation around an point of invasion) and wood (compartmentalization of infected tissue and callusing) of cedar trees are effective at halting further spread of Armillaria in host tissue. Belowground infection on both Douglas-fir and western red cedar is almost certainly higher than what was detected in the disease incidence surveys. However, based on our inoculation studies, it is more likely that the presence of such resistance mechanisms results in an infection being confined to tissue immediately surrounding a point of invasion, thereby limiting the extent of cambial invasion which might otherwise result in higher mortality rates in cedar.

22 CONCLUSION AND MANAGEMENT IMPLICATIONS: 22 The practical implications of using western red cedar are significant. In stands with moderate to high levels of infection, cedar will act as a barrier to disease spread between susceptible conifer species because it is most likely that an infection by Armillaria will induce formation of resistance mechanisms to confine the fungus to a lesion in the bark or wood. Resistance in one particular species has direct implications on adjacent crop trees such as reducing the amount of secondary inoculum which would otherwise be available to the fungus to infect new hosts. The results presented here make us confident that recommending the inclusion of western red cedar in higher proportions when planting infested sites may reduce the overall impact of Armillaria root disease.

23 23 REFERENCES: Blanchette, R.A, and A.R. Biggs Defense mechanisms of woody plants against fungi. Springer-Verlag, New York. 458 pp. Cleary, M; van der Kamp, B; and Morrison, D Host response to infection by Armillaria ostoyae in the roots of western red cedar and western hemlock in the southern interior of British Columbia. In: Proceedings of the 11 th IUFRO Root and Butt Rot Conference, Aug , Poznan and Bialowiesza, Poland. [In press] DeLong, D.L A Retrospective Investigation of Advanced Western Redcedar Regeneration in the ICHwk1, ICHmw2, and ICHmw1 of the Nelson Forest Region Experimental Project Res, Br., B.C. Min. For., Victoria, B.C. Work Pap. 25/1997. Morrison, D.J., and Pellow, K.W Development of Armillaria root disease in a 25- year-old Douglas-fir plantation. In Proceedings of the 8 th International Conference on Root and Butt Rots, Aug. 1993, Wik, Sweden, and Haikko, Finland. Edited by M. Johansson and J. Stenlid. Swedish University of Agricultural Sciences, Uppsala. pp Morrison, D.J., Pellow, K.W., Norris, D.J., and Nemec, A.F.L Visible versus actual incidence of Armillaria root disease in juvenile coniferous stands in the southern interior of British Columbia. Can. J. For. Res. 30: Morrison, D.J., Pellow, K.W., Nemec, A.F.L., Norris, D.J. and P. Semenoff The effects of partial cutting on the epidemiology of Armillaria root disease in the southern interior of British Columbia. Can. J. For. Res. 31: Reynolds, K.M, and Bloomberg, W.J Estimating probability of inter-tree root contact in second-growth Douglas-fir. Can. J. For. Res. 12: Robinson, R.M Response of western larch and Douglas-fir to infection by Armillaria ostoyae. Ph.D. Thesis, Faculty of Forestry, University of British Columbia, Vancouver, B.C. Robinson, R.M; Morrison, D.J Lesion formation in the roots of western larch (Larix occidentalis Nutt.) and Douglas-fir (Pseudotsuga menziesii) in response to infection by Armillaria ostoyae (Romagn.) Herink. For. Path. 31, Wahlstrom, K.T. and Johansson, M Structural responses in bark to mechanical wounding and Armillaria ostoyae infection in seedlings of Pinus sylvestris. Eur. J. For. Path. 22:65-76.

24 Appendix I 24 Total number of hardwood trees tallied by species in different disease status categories HARDWOOD SPECIES a DISEASE STATUS W Ac At Alder Maple Ep TOTAL Healthy Infected by A. ostoyae (PROGRESSIVE infection at the root collar) Infected by A. ostoyae (CALLUSED infection at the root collar) Dead (KILLED by A. ostoyae) Dead (unknown/other factors) Total a W = willow; Ac = cottonwood; At = trembling aspen; alder = red alder; maple = Douglas-maple; Ep = paper birch