Phytophthora ramorum Causes Cryptic Bole Cankers in Canyon Live Oak

Transcription

1 Plant Health Research Phytophthora ramorum Causes Cryptic Bole Cankers in Canyon Live Oak Tedmund J. Swiecki and Elizabeth A. Bernhardt, Phytosphere Research, Vacaville, CA 95687; Kamyar Aram and David M. Rizzo, Department of Plant Pathology, University of California, Davis 95616; Takao Kasuga and Mai Bui, Crops Pathology and Genetics Research Unit, USDA-ARS, Davis, CA Accepted for publication 12 February Published 18 February ABSTRACT Swiecki, T. J., Bernhardt, E. A., Aram, K., Rizzo, D. M., Kasuga, T., and Bui, M Phytophthora ramorum causes cryptic bole cankers in canyon live oak. Plant Health Prog. 17: Mortality of large canyon live oaks suddenly appeared in natural stands in San Mateo, CA, starting in A survey of affected stands showed that symptomatic trees were spatially associated with California bay, the primary source of Phytophthora ramorum spores in California coastal oak forests. Trunk canker symptoms on affected trees were similar to latestage symptoms caused by P. ramorum on other oak hosts, but the pathogen could not be isolated from affected trees. Artificial inoculation of logs, and later, trees, confirmed that P. ramorum caused phloem cankers on canyon live oak, but cankers showed either no or minuscule external bleeding. Knowledge of early bark symptom appearance facilitated successful isolations from naturally infected trees. Tree declines associated with similarly cryptic Phytophthora cankers could remain undiagnosed or misdiagnosed for many years, thwarting detection and management efforts. INTRODUCTION Canyon live oak (Quercus chrysolepis Liebm.) is a widely distributed native California oak in the intermediate oak subgenus (Griffin and Critchfield 1972). Around 2007, unexplained mortality of many large canyon live oaks became apparent in several locations in San Mateo Co., CA. Affected canyon live oaks had progressive canopy thinning and trunk cankers with fruiting bodies of Annulohypoxylon thouarsianum (Lév.) Ju, Rogers & Hsieh. The wood beneath many cankers was extensively colonized by flat-headed borers (Buprestidae) and ambrosia beetles (Monarthrum spp.), but the trunks of recently killed trees were sound beyond a narrow band of decayed outer sapwood. Measurements of midday stem water potential (McCutchen and Shackel 1992) of symptomatic and asymptomatic trees were similar and did not indicate trees were under severe water stress. No signs of root decay were seen in root crown inspections and the spatial distribution of affected trees did not suggest root diseases. Trunk cankers with A. thouarsianum sporulation and beetle boring are commonly observed in late stages of sudden oak death (SOD), caused by Phytophthora ramorum Werres, de Cock & Man in't Veld (Rizzo et al. 2002). However, SOD was only known to occur in tanoak (Notholithocarpus densiflorus [Hook. & Arn.] Manos, Cannon & S.H.Oh), and a few California native oaks in the red oak subgenus: coast live oak (Quercus agrifolia Née); Shreve oak (Q. parvula Greene var. shrevei [C.H. Mull.] Nixon); and California black oak (Q. kelloggii Newberry). The conspicuous bleeding trunk cankers typically seen in the early Corresponding author: T. Swiecki. phytosphere@phytosphere.com. doi: / PHP-RS The American Phytopathological Society stages of SOD were not seen in affected canyon live oaks. Though P. ramorum was previously shown to cause dieback in small (<2- cm diameter) branches of canyon live oak (Murphy and Rizzo 2003), no evidence of trunk infections had been previously reported. In most P. ramorum hosts, infections are similarly limited to leaves or small branches (Garbelotto et al. 2003). In California oak forests, leaf infections of California bay (Umbellularia californica [Hook. & Arn.] Nutt.) are the most important source of P. ramorum spores (Davidson et al. 2005) and infected oaks normally occur in close proximity to California bay (Kelly and Meentemeyer 2002; Swiecki and Bernhardt 2002; Metz et al. 2012). California bays with foliar symptoms of P. ramorum infection and SOD-killed coast live oaks (Q. agrifolia) were intermixed with symptomatic canyon live oaks at affected sites. Although the pathogen was isolated from California bay leaves and coast live oak cankers in the area, we were unable to recover P. ramorum from about 50 different symptomatic canyon live oaks sampled at multiple dates and locations. We initiated the studies described here to determine whether P. ramorum can cause trunk cankers in mature canyon live oaks and could be the cause of the observed tree mortality. SPATIAL RELATIONSHIP BETWEEN CALIFORNIA BAY AND SYMPTOMATIC CANYON LIVE OAK Methods. To determine the spatial relationship between symptomatic canyon live oak and California bay, we surveyed an affected stand in Los Trancos Open Space Preserve, San Mateo Co., CA, in March 2009 ( N; W). Canyon live oak, in a wide range of size classes, was common in the stand. Coast live oak was a minor component in most of the stand, and many of these trees had SOD symptoms. California bay was distributed irregularly throughout the stand at varying densities. Previous sampling had confirmed that P. ramorum was well established on California bay in the preserve (unpublished PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 20

2 data) and foliar symptoms of P. ramorum infection were common. For each canyon live oak within the surveyed area, we recorded trunk diameter at 1.37 m above grade (DBH), and disease symptoms (presence of cankers with A. thouarsianum sporulation or beetle boring). We used a 100 mw green laser attached to an angle gauge to establish a vertical line at the edge of California bay canopy that was closest to the oak trunk. We measured the horizontal distance from that line to the oak trunk using a laser rangefinder (Leica Disto A5). California bay canopy cover around each canyon live oak was visually estimated and recorded categorically as 0 = no bay cover, 1 = 1 to 25%, 2 = 26 to 50%, 3 = 51 to 75%, 4 = 76 to 100% in concentric circular zones that had radii of 2.5 m and 5 m from the oak trunk. We constructed logistic regression models (JMP statistical software, SAS Institute Inc., Cary, NC) using the presence of SOD symptoms and tree mortality as binary outcome variables. California bay cover ratings, distances, and tree DBH were used as predictors. Association of California bay with symptomatic canyon live oak. Among surveyed canyon live oaks, 68% had trunks within one meter of bay foliage, and none of the surveyed oak trunks had more than 9.8 m of clearance from bay foliage. Among symptomatic canyon live oaks, 45 of 48 had bay foliage within 0.5 m of the trunk. Symptoms were present on 75% of the canyon live oaks that had a California bay cover rating of 4 (more than 75% cover) within 2.5 m of the trunk (Table 1). In contrast, oaks that did not have California bay foliage within 5 m of the trunk were all asymptomatic (Table 1). Logistic regression models showed that the likelihood of SODlike canker symptoms or mortality in canyon live oak increased with increasing cover of California bay within either 2.5 or 5 m of the oak trunk and decreased with increasing bay foliage-oak trunk distance. The likelihood of symptoms on oak also increased with increasing DBH (Table 2), which ranged from 15 to 180 cm (mean 57 cm). Among bay-related variables, California bay cover within 5 m of the trunk was the best predictor of disease based on corrected Akaike Information Criterion (AICc) and the misclassification rate of the predicted outcome (Table 2). Nonetheless, overall model and predictor significance levels were unchanged if California bay cover within 2.5 m or minimum oak trunk-bay foliage clearance were substituted into the model. The three bayrelated variables were highly collinear and could not be combined in the same model. The survey included 14 dead canyon live oaks, all of which had oak trunk-bay canopy clearances of 0.5 m or less. Bay cover within 5 m was the strongest predictor of mortality, but the other bay variables could be substituted into the model without substantially changing model significance (P 0.001), AICc, or the misclassification rate. Oak DBH was not significant in the mortality models. These relationships are nearly identical to previously documented relationships between the local density and proximity of California bay and SOD symptoms in coast live oak stands (Swiecki and Bernhardt 2008). These relationships are related to inoculum density. The quantity of splash-dispersed P. ramorum inoculum is greatest near infected bay foliage and attenuates rapidly with increasing distance (Davidson et al. 2005; Swiecki and Bernhardt 2008). The effect of DBH may be related to tree size (increasing interception of splashed inoculum) or bark thickness, which is positively correlated with susceptibility in coast live oak (Swiecki and Bernhardt 2006). LOG INOCULATIONS Methods. To test whether P. ramorum could infect canyon live oak, we inoculated logs (trunk sections) cut from healthy coast live oak and canyon live oak trees growing in a forest on Bay cover rating (0 to 2.5 m) TABLE 1 Cross tabulation showing symptom status of canyon live oaks [symptomatic (bold)/total] by California bay cover ratings for 0 to 2.5 m and 0 to 5 m distance ranges (0 = no bay cover, 1 = 1 to 25%, 2 = 26 to 50%, 3 = 51 to 75%, 4 = 76 to 100%). Bay cover rating (0 to 5 m) Totals Overall percent symptomatic 0 0/12 3/11 3/23 13% 1 5/22 0/2 5/24 21% 2 0/2 6/11 4/5 10/18 56% 3 1/1 10/16 1/1 12/18 67% 4 18/24 18/24 75% Totals 0/12 8/35 7/14 14/21 19/25 48/107 Overall percent symptomatic 0% 23% 50% 67% 76% 45% Model AICc TABLE 2 Summary of three logistic regression models for presence of SOD-like symptoms (present/absent) in canyon live oak at Los Trancos Open Space Preserve using different predictors related to local California bay cover and proximity*. Mis-classification rate Model parameters Odds ratio Odds ratio 95% confidence interval Parameter likelihood ratio P level % Minimum oak-bay distance (m) < Oak DBH (cm) % CA bay cover within 2.5 m < Oak DBH (cm) % CA bay cover within 5 m < Oak DBH (cm) *Overall model P < , n = 105 for all models. Misclassification rate is calculated from the confusion matrix. PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 21

3 TABLE 3 California Phytophthora isolates used in inoculations. Phytophthora species Isolate x Host Experiment y County and collection year P. cinnamomi P-541 Arbutus menzesii L Santa Cruz (2009) P. ramorum Pr-500 Rhododendron sp. L Sacramento (2006) P. ramorum Pr-510 Rhododendron sp. L Sacramento (2006) P. ramorum Pr-52 Rhododendron sp. L Santa Cruz (2000) P. ramorum Pr-155 Notholithocarpus densiflorus L Santa Clara (2001) P. ramorum Pr-461 Quercus chrysolepis L Humboldt (2005) P. ramorum Pr-487 Umbellularia californica L Sonoma (2005) P. ramorum Pr-710 U. californica L, T1 San Mateo (2007) P. ramorum Pr-745 U. californica (rainwater run-off) T1 Santa Clara (2010) P. ramorum Pr-1556 U. californica T2 Santa Clara (2011) P. ramorum Pr-1557 U. californica T2 Santa Clara (2011) x All isolates maintained in the collection of D. M. Rizzo, Univ. of California, Davis. Isolates Pr-500 and Pr-510 are NA-2 lineage, all others NA-1. y L = used in log inoculation experiment, T1 = used in first tree inoculation experiment, T2 = used in reinoculation tree experiment. Midpeninsula Regional Open Space District land near Woodside, CA. Individual logs (average 1.3 m long and 0.13 m diameter) were cut from each of nine canyon live oak and nine coast live oak trees in late August The cut ends were painted with a sealant (Waxlor; Willamette Valley Company, Eugene, OR) immediately after cutting. Logs were moved to a growth chamber maintained at 20 C and were inoculated within 2 days after collection. Each of the 18 logs was inoculated with seven different P. ramorum isolates, one P. cinnamomi isolate (Table 3), and one mock inoculation with a sterile agar plug. Following published methods (Brasier and Kirk 2001; Hansen et al. 2005), a 9-mm-diameter plug was cut from an actively growing colony on clarified V8-juice agar [66 ml of V8 juice (Campbell Soup Co., Camden, NJ) neutralized with CaCO 3 and clarified by centrifuging, and 17 g of agar/liter] and placed into a hole made through the bark to the cambium with a 9-mm diameter (#6) cork borer. The bark plug was replaced and the inoculation point covered with sterile cotton wetted with sterile water. A piece of aluminum foil was taped over the inoculation site and the log was wrapped with Parafilm to cover the aluminum foil. Inoculated logs were sprayed with sterile water, sealed in plastic bags, and placed in a growth chamber at 20 C. Logs were held for 40 to 49 days, at which point the outer bark was removed and the inner bark was examined for cankers. One log from each species was evaluated after 64 days. The outer margins of cankers were traced onto clear plastic film. Canker areas were determined from scans of the traced outlines using Assess software (American Phytopathological Society). For very large and symmetrical cankers, mostly occurring on Q. agrifolia, the canker area was estimated as the area of an ellipse with canker length and width as axes. For all inoculations, pieces of bark were taken from the edges of discolored phloem at the cardinal direction points and placed in PARP semiselective agar media (Erwin and Ribeiro 1996) in petri plates. Plates were incubated at 18 C. Canker development on inoculated logs. Bleeding did not occur on inoculated canyon live oak logs and rarely occurred on coast live oak logs. Cankers developed in the phloem of inoculated logs by 6 weeks after inoculation. Only a few inoculations (3 of 63 for both canyon live and coast live oak with P. ramorum, 2 of 9 for coast live oak with P. cinnamomi) failed to produce more discoloration than the agar controls; these unsuccessful inoculations were excluded from canker size analyses. For P. ramorum, all canker size parameters (Table 4) were significantly smaller in canyon live oak than in coast live oak (t-test P < ). All canker size parameters for P. cinnamomi on canyon live oak were not significantly different from those of P. ramorum. Area, length, and width of P. Variable TABLE 4 Canker size and bark thickness (± standard deviation) of canyon live oak and coast live oak logs inoculated with P. ramorum*. Canyon live oak Coast live oak Mean canker area (cm 2 ) 61.7 ± ±168.2 Mean canker length (cm) 17.5 ± ± 8.3 Mean canker width (cm) 3.9 ± ± 5.8 Mean canker length:width ratio 4.73 ± ± 0.62 Mean bark thickness (% of log radius) 8.0 ± ± 4 *All parameters differ significantly between oak species at P < (t test). Sample sizes: canyon live oak n = 60; coast live oak n = 59 for cankers; n = 9 for both oak species for bark thickness. cinnamomi cankers were all considerably smaller than those of P. ramorum on coast live oak (t test P = , 0.021, and respectively), although length:width ratio did not differ. Cankers in canyon live oak were narrower overall (higher length:width ratio) than those in coast live oak (Table 4). This may be related to the much higher density of thick-walled vertical fibers found in the bark of canyon live oak compared with coast live oak (unpublished data). Based on the ratio of bark thickness to log radius, canyon live oak bark was about half the thickness of coast live oak bark (Table 4). This thinner bark may also slow the radial spread of cankers. P. ramorum was readily isolated from the canker margins of all inoculations in both oak species (77% average recovery per canker) but was not recovered from the mock inoculations. Percent recovery of P. cinnamomi from cankers did not differ by oak species or from that of P. ramorum. The log inoculations indicated that P. ramorum could infect bark of large canyon live oak stems and provided the first definitive information about the internal appearance of early stage cankers. However, detached logs did not reliably develop the same range of external symptoms that developed on intact trees, such as bleeding. TREE INOCULATIONS Inoculation methods. In July 2010, we initiated a field test by inoculating nine canyon live oaks (25 cm average DBH, range 23 to 31 cm) in each of two plots in adjacent stands at Skyline Ridge Open Space Preserve, San Mateo Co., CA ( N; W). The lower plot area was a closed-canopy stand dominated by relatively tall (18 to 21 m) canyon live oaks with small crowns. Some of these trees were partially overtopped and PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 22

4 had thinning canopies. The upper plot area was in a 25-year-old closed-canopy stand that had been planted from locally collected seed. Trees in this stand were much shorter (8 to 12 m) with wider, denser crowns. Trees were inoculated with 1-cm-diameter agar plugs taken from the margins of actively growing P. ramorum cultures following procedures previously used for coast live oak (Rizzo et al. 2002). Each tree was inoculated as described above for log inoculations with two different P. ramorum isolates (Table 3) recovered locally and a control mock inoculation (sterile agar only). Inoculation points were 1.5 m above ground level on average and were evenly spaced around the circumference of the tree. Two Shreve oak trees growing in the upper plot were inoculated to serve as positive controls. After inoculation, symptoms were evaluated six times over the first year and less frequently up to 59 months after inoculation. After 5 months, cankers of four canyon live oaks (two trees per plot) and one Shreve oak were exposed by removing the outer bark around the inoculation points with a draw knife. Trees selected in this initial sample included those with the most conspicuous external symptoms during the previous 5 months. Cankers of five additional canyon live oaks (three from lower plot, two from upper plot) were exposed after 9 months, and cankers of four additional trees (two in each plot) were exposed after 2 years. Canker margins were traced on clear plastic film. ImageJ software (U.S. National Institutes of Health, Bethesda, MD) was used to measure canker areas from digital scans of the tracings. Canker areas were log transformed prior to data analysis because of the wide range in canker areas. Transformed canker areas were analyzed using analysis of variance and matched pairs platforms in JMP statistical software (SAS Institute Inc.). In July 2012, we reinoculated four of the canyon live oaks with two other local isolates (Table 3). One highly susceptible tree in the upper plot had two trunks. The healthy, previously noninoculated trunk was used for the reinoculation. Other trees showing highly susceptible reactions could not be used because existing cankers did not leave enough healthy bark on the trunks for reinoculation. Hence, three of the four reinoculations were made on trees that had developed relatively small cankers. After 5 months, cankers from the reinoculations were exposed by removing the bark around the inoculation points. Isolations were made from the edges of canker margins as described below. Pathogen recovery methods. Multiple tissue pieces were taken from the edges of the exposed cankers and placed into PARP medium (Fig. 1). The number of sample pieces collected increased with the size of the canker and ranged from 4 to 45. To determine whether isolations were more effective from particular portions of the canker, sampled areas were numbered, marked, and recorded via digital photographs to document sampling locations. In most cankers, we sampled at two depths. Most samples were collected in the mid to outer live phloem tissue. A second set of samples were taken by cutting deeper to the cambium and outer sapwood and sampling the margin of tissue discoloration at the greatest depth from the bark surface. External symptom development on inoculated trees. Few external symptoms developed on inoculated canyon live oaks. By 8 months after inoculation, some bleeding developed around 7 of 36 (19%) inoculation points on 6 of 18 canyon live oaks. The amount of bleeding was usually limited to a few drops 1 to 2 mm in diameter. Only one tree developed a canker with easily visible amounts of bleeding. This tree and one other had active bleeding for periods of up to 3 months, whereas on others active bleeding was visible for shorter periods. In comparison, all four P. ramorum inoculation points on the two Shreve oaks developed cankers with conspicuous bleeding by 8 months. One Shreve oak had active bleeding for at least 3 months after this point, until beetles invaded. Bleeding was observed in the other Shreve oak up to 46 months after inoculation. Ambrosia beetle boring was observed 11 months after inoculation in one canyon live oak that never developed bleeding. A. thouarsianum sporulation developed on this same tree by 17 months after inoculation and was seen on one additional tree 40 FIGURE 1 Appearance of exposed inner bark tissues around inoculation points (red circles, each 1 cm diameter) on canyon live oak: (top) mock inoculation at 5 months; (bottom) a large canker sampled 5 months after inoculation. P. ramorum was recovered by isolation on PARP medium from sample points indicated by white pointers; the pathogen was not recovered at sample points circled in yellow. PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 23

5 months after inoculation. Through 59 months of observation, none of the inoculated canyon live oaks died, even though several were nearly girdled. One tree was almost completely girdled when sampled 5 months after inoculation. This tree was still alive 59 months after inoculation, although the canopy was thinning and had curled, off-color foliage. In comparison, one of the Shreve oaks had ambrosia beetles by 17 months after inoculation, and was dead, with foliage turned brown and dry, by 29 months. The lower trunk of the other Shreve oak tree had recent ambrosia beetle boring around its entire circumference 59 months after inoculation, and had a green canopy despite substantial (>50%) dieback. By 20 months after inoculation, some of the canyon live oaks had developed external bark cracking above and below inoculation points, which was associated with callus formation around small cankers. Callus development was seen at 15 of 36 (42%) inoculation points at 23 months and at 23 of 36 (64%) inoculation points at 40 months. Mock inoculation wounds in some trees were completely callused over after 23 months. The Shreve oak that was still green 59 weeks after inoculation showed some callus development along one canker edge at 40 months post-inoculation, but the canker continued to expand and active bleeding was seen at 46 months. Canker development on inoculated trees. Cankers were seen at all sampled P. ramorum inoculation points on canyon live oak whether or not external symptoms were present (Fig. 1). Cankers around P. ramorum inoculation points varied widely in size (Fig. 2), ranging from 9 to 1567 cm 2, with a mean of 350 cm 2. Canker lengths ranged from 5 to 69 cm, with a mean of 27 cm. Necrosis around mock inoculations ranged from 2 to 18 cm 2, with a mean of 5 cm 2 (lengths 2 to 9 cm, mean 3 cm). Necrotic areas were significantly smaller for mock inoculations than P. ramorum inoculations (ANOVA contrast P < ), with no difference in size between the two P. ramorum isolates. Cankers formed by the two isolates on a given tree were similar in size (Fig. 2) and did not differ significantly in a matched pairs analysis by tree. Canker length:width ratios did not differ by sample date and averaged 3.5 ± 1.8. Canker size varied widely between individual trees in both plots, but the largest cankers were seen on trees in the upper plot, which had younger, faster-growing trees. Similarly, in coast live oak, more dominant, fast-growing trees exhibit greater susceptibility to SOD than slow-growing, suppressed trees (Swiecki and Bernhardt 2006). The four canyon live oaks that were reinoculated in 2012 developed similar size cankers to those from the first inoculation. One tree that showed a highly susceptible reaction (mean canker area 1469 cm 2 ) developed similarly large cankers when its second trunk was inoculated (mean 1323 cm 2 ), whereas the three relatively resistant trees again produced small cankers (mean canker areas 13 to 63 cm 2 across both inoculations). This suggests that canker size was influenced by the genotype and/or physiological condition of the host and that prior infection did not significantly influence the disease progression of the second infection. Pathogen recovery from inoculated trees. P. ramorum was reisolated from all cankers that formed at eight inoculation points on the four canyon live oaks sampled at 5 months. Recovery of the pathogen from tissue pieces around individual cankers ranged from 14% to 75%, with an overall average of 48% per canker. One Shreve oak was also sampled at 5 months; P. ramorum recovery from its two cankers was 81% and 95%. At 9 months, the pathogen was successfully reisolated from nine of 10 sampled canyon live oak cankers (20% to 83% recovery from individual cankers) at an overall rate of 43%. At 24 months after inoculation, P. ramorum was not reisolated from cankers that formed at eight inoculation points on four sampled canyon live oaks. Cankers on these trees were all relatively small, and were either limited by callus or appeared inactive. When the four reinoculated trees were sampled 5 months after reinoculation, P. ramorum was reisolated FIGURE 2 Canker areas (cm 2 ) from individual canyon live oaks inoculated with two different Phytophthora ramorum isolates (745 and 710) and a control (mock inoculation) sampled at varying times after inoculation, grouped by plot. Trees that showed the most bleeding around cankers were sampled first. PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 24

6 from six of the eight cankers and 31% of tissue pieces yielded P. ramorum (14% to 70% recovery per canker, excluding the two cankers with no recovery). All isolations from mock inoculations were negative at all sample dates. Among 12 cankers sampled at 5 or 9 months with both phloem and xylem isolations, the isolation rate from the xylem (overall 32% recovery) was not significantly different from that of the phloem (overall 44% recovery) (2-tailed paired t test). Examination of recovery from marked spots on the cankers did not reveal any consistent visible difference between spots that yielded positive or negative results (Fig. 1). Plots did not differ with respect to recovery of P. ramorum from cankers at either sampling date. PATHOGEN RECOVERY FROM NATURALLY INFECTED TREES Insights gained from the field inoculation improved our ability to locate and recognize early-stage P. ramorum cankers in canyon live oak. Isolation efficiency data indicated that isolation from active cankers has a high likelihood of yielding P. ramorum. To date, we have isolated P. ramorum from trunk cankers in six naturally infected mature canyon live oaks in native stands in San Mateo County. All six trees were adjacent to, or overtopped by, California bay. Cankers in most of these trees were located by observing 1- to 2-mm-diameter drops of dark ooze on the bark (Fig. 3), visible only under careful inspection with a highintensity flashlight. One symptomatic tree only showed dark staining on the bark, probably from spreading of ooze by rain. Chipping away the outer bark revealed cankers similar in appearance to those obtained through artificial inoculation (Fig. 3). Several of the cankers extended 10 to 20 cm or more beyond the area where bleeding was observed. Over the next 2 years, one of these infected trees showed canopy thinning; the canker expanded in size and was colonized by A. thouarsianum and ambrosia beetles. Another tree was already heavily cankered with extensive canopy thinning, and both canker extent and canopy thinning have progressed. Cankers in two trees showed no evidence of expansion or invasion by secondary organisms by 3 to 4 years after the positive isolations were obtained. CONCLUSIONS Based on the spatial relationship between symptomatic canyon live oak and California bay, inoculation experiments on logs and mature trees, and isolations from naturally infected trees, we conclude that P. ramorum causes trunk cankers in canyon live oak. Canyon live oak was not among the native California oaks that were identified as canker hosts of P. ramorum shortly after the pathogen was discovered in 2000 (Rizzo et al. 2002; Rizzo et al. 2005). More than 10 years elapsed before we could confirm that mature canyon live oak is also susceptible to lethal P. ramorum trunk cankers (Aram et al. 2011; Swiecki et al. 2013). One reason for this is that forests containing both California bay and hosts such as coast live oak are much more common and widely distributed than forests containing both California bay and canyon live oak (Griffin and Critchfield 1972). Also, SOD symptoms on mature canyon live oak only started to become common after successive years favorable for disease development occurred in 2005 and 2006, whereas SOD symptoms appeared in many coast live oak stands before However, diagnosis of P. ramorum infection in canyon live oaks was mainly impeded by the lack of external symptoms on recently infected canyon live oaks. We observed only limited and sporadic bleeding in both field-inoculated and naturally infected canyon live oaks. In contrast, bleeding cankers are one of the primary early symptoms of trunk infections caused by P. ramorum in tanoak and susceptible species in the red oak subgenus (Rizzo et al. 2002; Rizzo et al. 2005). When the outer bark is sliced away, early-stage cankers in canyon live oak resemble those reported by Rizzo et al. (2002) for coast live oak, but these early-stage cankers are rarely detectable due to a lack of visible external symptoms. We have isolated P. ramorum from only a few naturally infected canyon live oaks, all of which showed minute amounts of bleeding. Cankers in cryptically infected canyon live oak trees are generally not visible until they are colonized by secondary organisms, at which point P. ramorum cannot be recovered. A significant implication of our results is that lethal trunk cankers caused by P. ramorum or other Phytophthora species may be very difficult to detect in woody hosts that develop little or no bleeding and in which mortality occurs years after infection. Tree mortality associated with canker-causing Phytophthora species that are very difficult to recover is likely to be attributed to other agents or categorized as a decline of unknown cause. The recognition that many large canyon live oaks were dying or at risk of being killed by P. ramorum cankers has led to ongoing management studies designed to prevent P. ramorum infection in canyon live oaks. The inoculation tests and risk models related to California bay proximity suggest that canyon live oak may be more resistant to infection than coast live oak (Swiecki and Bernhardt 2008) and Shreve oak. Even after five years, none of the 18 canyon live oaks inoculated in the field have died, even though extensive stem cankers have developed in some of the trees. In contrast, Rizzo et al. (2002) reported that among five inoculated coast live oaks, two had died and one was nearly dead 13 months after inoculation. Furthermore, field inoculations provide evidence that individual canyon live oaks vary widely in susceptibility to P. ramorum. Either active or passive selection for FIGURE 3 Naturally infected canyon live oak from which Phytophthora ramorum was isolated in December The dark brown canker has been exposed by chipping away the outer bark. A small amount of bleeding (arrows) is visible below the chipped area. Foliage of an adjacent California bay is visible on the right side of the image. PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 25

7 resistant genotypes may be possible for long-term disease management (Hayden et al. 2011, 2013; Nagle et al. 2011). Managing the disease by reducing inoculum loads via removal of nearby California bay (Swiecki et al. 2013) should be more effective for canyon live oak than for other susceptible oaks. However, the delayed appearance of external canker symptoms and tree mortality complicates management of SOD-affected canyon live oak stands. Because the pretreatment disease status of subject trees cannot be reliably determined, it is difficult to assess efficacy of preventative treatments. Although log inoculations did not provide useful information on external symptom development, our results further validate the use of the technique as a screening tool for assessing the susceptibility of tree species to trunk diseases caused by Phytophthora species (e.g., Brasier and Kirk 2001; Hansen et al. 2005; Ireland et al. 2012). The canyon live oak logs we inoculated developed substantially larger lesions (Table 3) than the average 11-cm 2 lesions reported by Hansen et al. (2005) from two inoculated canyon live oak logs. Overall canker sizes in our canyon live oak log and tree inoculations were similar to those seen for moderately to highly susceptible species in other studies (Brasier and Kirk 2001; Rizzo et al. 2002; Ireland et al. 2012). ACKNOWLEDGMENTS This research was supported by the USDA Forest Service, State and Private Forestry, the Gordon and Betty Moore Foundation, and the Midpeninsula Regional Open Space District. We thank Cindy Roessler and Alex Hapke for assistance with site selection and monitoring, Kylie Ireland, Akiko Oguchi, Mike Jones, and Juan Moral for assistance with log evaluations. LITERATURE CITED Aram, K., Swiecki, T., Bernhardt, E., and Rizzo, D. M Canyon live oak (Quercus chrysolepis) is susceptible to bole infection by Phytophthora ramorum. Phytopathology 101:S8. Brasier, C. M., and Kirk, S. A Comparative aggressiveness of standard and variant hybrid alder phytophthoras, Phytophthora cambivora and other Phytophthora species on bark of Alnus, Quercus and other woody hosts. Plant Pathol. 50: Davidson, J. M., Wickland, A. C., Patterson, H. A., Falk, K. R., and Rizzo, D. M Transmission of Phytophthora ramorum in mixed-evergreen forest in California. Phytopathology 95: Erwin, D. C., and Ribeiro, O. K Phytophthora Diseases Worldwide. American Phytopathological Society, St. Paul, MN. Garbelotto, M., Davidson, J. M., Ivors, K., Maloney, P. E., Hüberli, D., and Rizzo, D. M Non-oak native plants are the main hosts for the sudden oak death pathogen in California. Cal. Agric. 57: Griffin, J. R., and Critchfield, W. B The distribution of forest trees in California. Res. Paper PSW-RP-82. Pacific Southwest Forest and Range Exp. Stn., Forest Service, USDA, Berkeley, CA. Hansen, E. M., Parke, J. L., and Sutton, W Susceptibility of Oregon forest trees and shrubs to Phytophthora ramorum : A comparison of artificial inoculation and natural infection. Plant Dis. 89: Hayden, K. J., Garbelotto, M., Dodd, R., and Wright, J. W Scaling up from greenhouse resistance to fitness in the field for a host of an emerging forest disease. Evol. Appl. 6: Hayden, K. J., Nettel, A., Dodd, R. S., and Garbelotto, M Will all the trees fall? Variable resistance to an introduced forest disease in a highly susceptible host. For. Ecol. Manag. 261: Ireland, K. B., Hüberli, D., Dell, B., Smith, I. W., Rizzo, D. M., and Hardy, G. E. S. J Potential susceptibility of Australian native plant species to branch dieback and bole canker diseases caused by Phytophthora ramorum. Plant Pathol. 61: Kelly, M., and Meentemeyer, R. K Landscape dynamics of the spread of sudden oak death. Photogramm. Eng. Remote Sens. 68: McCutchen, H., and Shackel, K. A Stem-water potential as a sensitive indicator of water stress in prune trees (Prunus domestica L. cv. French). J. Am. Soc. Hortic. Sci. 117: Metz, M. R., Frangioso, K. M., Wickland, A. C., Meentemeyer, R. K., and Rizzo, D. M An emergent disease causes directional changes in forest species composition in coastal California. Ecosphere 3:art86. Murphy, S. K., and Rizzo, D. M First report of Phytophthora ramorum on canyon live oak in California. Plant Dis. 87: Nagle, A. M., Mcpherson, B. A., Wood, D. L., Garbelotto, M., and Bonello, P Relationship between field resistance to Phytophthora ramorum and constitutive phenolic chemistry of coast live oak. For. Pathol. 41: Rizzo, D. M., Garbelotto, M., Davidson, J. M., Slaughter, G. W., and Koike, S. T Phytophthora ramorum as the cause of extensive mortality of Quercus spp. and Lithocarpus densiflorus in California. Plant Dis. 86: Rizzo, D. M., Garbelotto, M., and Hansen, E. M Phytophthora ramorum: Integrative research and management of an emerging pathogen in California and Oregon forests. Annu. Rev. Phytopathol. 43: Swiecki, T. J., and Bernhardt, E. A Increasing distance from California bay reduces the risk and severity of Phytophthora ramorum canker in coast live oak. Pages in: Proc. of the Sudden Oak Death 3rd Science Symp. PSW-GTR-214. Pacific Southwest Res. Stn., Forest Service, USDA, Albany, CA. Swiecki, T. J., Bernhardt, E., Aram, K., and Rizzo, D Diagnosis and management of Phytophthora ramorum in canyon live oak, an atypical bole canker host. Pages in: Proc. of the Sudden Oak Death 5th Science Symp.. Gen. Tech. Rep. PSW-GTR-243. Pacific Southwest Res. Stn., Forest Service, USDA, Albany, CA. Swiecki, T. J., and Bernhardt, E Evaluation of stem water potential and other tree and stand variables as risk factors for Phytophthora ramorum canker development in coast live oak. Pages in: Proc. of the 5th Symp. on California Oak Woodlands. PSW-GTR-184. Pacific Southwest Res. Stn., Forest Service, USDA, Albany, CA. Swiecki, T. J., and Bernhardt, E Disease risk factors and disease progress in coast live oak and tanoak affected by Phytophthora ramorum canker (sudden oak death). Pages in: Proc. of the Sudden Oak Death Second Symp.: The State of Our Knowledge. PSW-GTR-196. Pacific Southwest Res. Stn., Forest Service, USDA, Albany, CA. PLANT HEALTH PROGRESS Vol. 17, No. 1, 2016 Page 26