AN ABSTRACT OF THE THESIS OF

Transcription

1 AN ABSTRACT OF THE THESIS OF Amy I. Eckert for the degree of Master of Science in Forest Science presented on March 15, 007. Title: Field Classification of White Pine Blister Rust Stem-Cankers on Resistant Western White Pine in Northern Idaho and Determination of Respective Tissue Damage Through Tree Ring Analysis Abstract Approved: Gregory M. Filip Western white pine historically dominated northern Idaho s forested landscape and was the Inland Empire s most economically important tree. White pine blister rust, caused by the exotic fungus Cronartium ribicola, played a principal role in the decline of western white pine. The pathogen causes branch and bole cankers, which usually girdle and kill their host. Efforts to restore western white pine populations are underway. Despite gains in resistance during over 50 years of tree improvement efforts, high infection rates still challenge growth and management of rust-resistant western white pine in northern Idaho. One rust- resistance expression of particular promise in tree breeding orchards is abnormal or slow stem-canker growth. Abnormal or slow stem-canker growth also appears to occur in pole-sized (15-0 years old) plantations of rust resistant western white pine in northern Idaho. The extent to which outward stem canker appearance is related to internal tree tissue damage has not been previously addressed.

2 In this study, between 004 and 006, blister rust-caused stem cankers were grouped into three classes of expected severity based on field assessment of exterior canker characteristics. Cankers with the most aggressive looking characteristics (those with bright orange diamond-shaped margins, heavy resinosous, and no apparent defensive reaction of the host tree) were classified as Class I cankers. Cankers that had limited or no orange margins and had mild to moderate resinosous were classified as Class II cankers. Cankers with the least aggressive looking characteristics (those with no orange margins, tree tissue swelling to the extent that the swelling made a vertical depression on the tree bole, no resinosous, and an inactive appearance) were classified as Class III cankers. Stem cankers were examined on pole-sized, rust-resistant trees in three western white pine plantations in northern Idaho. After field classification, the cankers were harvested and cross-sectioned to determine: (1) the extent to which exterior canker appearance is related to internal tree tissue damage, () if tree circumference just prior to canker initiation is related to tree tissue damage during the first year following canker initiation, and (3) the relative impact of different class cankers on tree growth. Measurements of cross-sectioned cankers indicate statistically significant differences in tree circumferential tissue damage (percent tree girdle) during years 1 through 5 following canker initiation between canker Classes I and III, and between canker Classes II and III during years 1 through 8. Difference in median percent girdle was not significant between Classes I and II. Additionally, tree circumference just prior to canker initiation and tree tissue damage during the first year following canker

3 initiation were positively correlated; as tree size increased, first year percent girdle of host trees also increased. Finally, the ratio of circumferential tree growth post-canker initiation to growth pre-canker initiation differed significantly among trees infected with different class cankers. Differences were significant between canker Classes I and III and between Classes II and III, but not between Classes I and II. Growth ratios were smaller for trees with more aggressive appearing cankers. These research results will help foresters field-classify stem cankers and predict performance of infected rust-resistant western white pine. Knowing the relative tree damage and impact on tree growth of cankers with different exterior characteristics will help foresters develop silvicultural prescriptions, including prescriptions for pruning and thinning to encourage development of target stand composition and structure. Abnormal or slow-growing cankers may indicate that their host will survive longer than expected. In addition, observed differences in canker morphology and growth of infected trees may enhance tree breeders and geneticists investigations of slow canker growth mechanisms.

4 Copyright by Amy I. Eckert March 15, 007 All Rights Reserved

5 Field Classification of White Pine Blister Rust Stem-Cankers on Resistant Western White Pine in Northern Idaho and Determination of Respective Tissue Damage through Tree Ring Analysis by Amy I. Eckert A THESIS submitted to Oregon State University in partial fulfillment of the requirements for the degree of Master of Science Presented March 15, 007 Commencement June 007

6 Master of Science thesis of Amy I. Eckert presented on March 15, 007. APPROVED: Major Professor, representing Forest Science Head of the Department of Forest Science Dean of the Graduate School I understand that my thesis will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my thesis to any reader upon request. Amy I. Eckert, Author

7 ACKNOWLEDGEMENTS I sincerely appreciate my major professor Dr. Greg Filip. I would not have started a Master s project without his acceptance and encouragement. I deeply appreciate the great coaching and guidance from my minor professor Dr. Mike Newton. I am also extremely grateful to Dr. John Schwandt for writing the proposal for funding provided by the USDA Forest Service Special Technology Development Program. Thankfully, Jeff Stone s door was always open to me for pathology discussion and questions. I owe my interest and educational foundation in forest fungal pathology to Dr. Stone s teaching and mentoring. I also owe many thanks to Brennan Ferguson for passing along some hard-to-find literature as well as some good blister rust discussions. I appreciate the guidance of Dr. Jim Byler, who came out of retirement and into the field with me in order to stretch my mind and guide my thinking about blister rust cankers. I am very grateful for the field help provided by Doug Wulff, Gary Kempton and Tom Eckberg. My family also played an important support role in this project. My mother is the best field assistant I ve ever met, and my father helped me keep my ducks in a row. Words cannot capture the love and gratitude I feel for my husband. Finally, I thank God for my Christian church families in Corvallis, OR, St. Maries, ID, and Lewistown, MT who supported this project with prayers and buoyed me with friendship and love.

8 TABLE OF CONTENTS 1 Introduction Biology of C. ribicola. Page Pycniospores Aeciospores Urediniospores Teliospores Basidiospores Spread of White Pine Blister Rust North American Introduction of C. ribicola Intensification of White Pine Blister Rust Wave Years Western Spread of C. ribicola Economic and Ecological Impacts of Blister Rust Blister Rust Control Through Ribes spp. Eradication Resistance Breeding Program Slow Canker Growth Resistance Trait. 15. Thesis Research Direction Materials and Methods Canker Class Definition Class I Cankers Class II Cankers Class III Cankers Site Selection and Description. 3

9 TABLE OF CONTENTS (Continued) Page 3.3 Canker Selection Exterior Canker Measurements Canker Harvest Canker Cross Sectioning Photographing Cross Sections Analyzing Cross Sections Measuring Annual Ring Circumferential Length and Area Measuring Circumferential Tissue Damage Statistical Analyses Relationship between Annual Percent Girdle and Canker Class Method Used in Models A and B to Estimate and Compare Annual Percent Canker Girdle Over Time Across Canker Classes Model A: Estimating Percent Canker Girdle Among Classes I, II and III over Canker Ages 1 through Model B: Estimating Percent Canker Girdle Between Classes II and III over Canker Ages 1 through Relationship of Tree Circumference and Amount of Circumferential Tissue Damage at Stem Canker Initiation Relative Canker Impact on Tree Growth Across Canker Classes Results... 45

10 TABLE OF CONTENTS (Continued) Page 4.1 Annual Percent Canker Girdling Differs among Canker Classes I, II and III Tree Circumference is Positively Correlated with Tissue Damage Amount During Canker Initiation Canker Class Affects Relative Tree Growth Discussion Other Statistical Methods Considered Annual Percent Canker Girdling Differed between Canker Classes I and III, but not I and II Tissue Damage During Stem Canker Initiation Increased as Tree Circumference Increased Trees with Class I Cankers Grew Less Four Years After Stem Canker Initiation than Trees with Class III Cankers Management Implications Future Work Conclusions Literature Cited.. 73

11 LIST OF FIGURES Figure Page 1. Blister rust stem canker with typical morphological characteristics. 3. Class I canker with typical characteristics: no marginal swelling or obvious tree reaction, orange mycelial margin forms diamond shape around branch of infection entry Class I canker without orange margin, but obviously active around entry branch with marginal swelling or apparent tree reaction Class II canker with obvious swelling around entry branch and recent resin 1 5. Class III canker that created narrow depression in stem with no signs of fungal activity Stem sections cut from cankers on a 7.5 hp bandsaw were 5 cm thick Digital photograph of a 5 cm thick stem section from a standard fixed height Digital photograph of cankered stem section showing traced annual ring length, ring area, and circumferential tissue damage 8 9. Two models were necessary because Class I cankers older than 5 years were not analyzed Model A assumptions of normality were met for log 10 transformed circumferential infection percent (% girdling) Model A assumptions of homogeneous variance were met for log 10 transformed circumferential infection percent (%girdling) Model B assumptions of normality were met for log 10 transformed circumferential infection percent (%girdling) Model B assumptions of homogeneous variance were met for log 10 transformed circumferential infection percent (%girdling).. 35

12 LIST OF FIGURES (Continued) Figure Page 14. Assumptions of homogeneous variance were met for log 10 transformed circumferential tissue damage at canker age Normality assumptions were met for log transformed circumferential tissue damage at canker age Cross-section of Class II canker showing ring area of four-year tree growth pre-canker initiation (dots) and ring area of four-year tree growth post- canker initiation (bars) Assumptions of normality were met for log 10 transformed growth ratio (post infection/pre infection tree growth) Assumptions of homogeneous variance were met for log 10 transformed growth ratio (post infection/pre infection tree growth) Annual median percent girdle over time showing overlap of the means for Class I and Class II cankers Percent canker girdling of individual Class I cankers with corresponding empirical interquartile range Percent canker girdling of individual Class II cankers with corresponding empirical interquartile range. 51. Percent canker girdling of individual Class III cankers with corresponding empirical interquartile range Percent canker girdling of individual Class I cankers contrasted with Class II empirical interquartile range Percent canker girdling of individual Class II cankers contrasted with Class III interquartile range.. 53

13 LIST OF FIGURES (Continued) Figure Page 5. Percent canker girdling of individual Class III cankers contrasted with Class I interquartile range Scatterplot of actual and predicted values of infection amount at canker age 1 as tree circumference increases Scatterplot of actual and predicted values of percent canker girdle at canker age 1 as tree circumference increases Trees with Class I cankers grew less four years after stem infection than trees with Class III cankers... 59

14 LIST OF TABLES Table Page 1. The 5 spore types of C. ribicola vary in relative hardiness and traveling distance. 4. The causal pathogen of white pine blister rust, C. ribicola, has affected 8 of the 9 U.S. native species of white pine Description of 3 plantations from which cankers were harvested Akaike s Information Criterion was used to select the covariance matrix that best supports the data. UN(5) and TOEP (7) were selected to address the covariance patterns in Models A and B, respectively Type 3 test of fixed effects for Models A and B Estimated median percent girdle for Models A and B Estimates of differences in median percent girdle over time among canker classes A positive, multiplicative relationship between tree circumference and circumferential tissue damage was established Canker Class significantly affects relative tree growth ratio at any tree age Trees with Class I cankers grew less four years after stem-canker initiation than trees infected with cankers in other classes Differences in estimated growth ratios are detectable between canker Classes I and III and between canker Classes II and III, but not between canker classes I and II... 58

15 Field Classification of White Pine Blister Rust Stem-Cankers on Resistant Western White Pine in Northern Idaho and Determination of Respective Tissue Damage Through Tree Ring Analysis 1. Introduction Forest pathologists in Idaho have recently noticed that some white pine blister rust stem cankers (caused by Cronartium ribicola J.C. Fisch.) on rust-resistant western white pine depart from expected canker shape and behavior. Typical blister rust stem cankers look like diamond-shaped orange bark discolorations centered around the point of infection entry (usually a branch or needle bundle). Abnormal stem cankers depart from the typical canker appearance in shape and size and range from small round depressions to long, vertical crevices in the stem. This research grouped blister rust stem cankers from 3 northern Idaho rustresistant western white pine plantations into 3 classes of expected severity based on canker appearance. Tree rings and annual host tissue damage were then measured on 11 cross-sectioned cankers to determine (1) percent stem girdling over time by each class, () whether tree size affects amount of stem girdling during the first year after stem canker initiation, and (3) if canker class affects relative tree growth. Results will help familiarize foresters with variations in canker morphology, and provide new information for assessing western white pine field performance. When considering the practical application of this research, it is important to understand the basic biology and history of the disease. For instance, knowledge of the life cycle of C. ribicola (including spore stages and mode of infection helps the

16 observer understand the developmental stages of infection and canker initiation. Information about the introduction and spread of blister rust in North America also provides the necessary background from which managers can understand the mechanism of blister rust spread across plantations and forests. Additionally, outlining the economic and ecological impacts of blister rust disease in the west helps frame the present need for western white pine restoration. Finally, information on the rustresistance breeding program, including the slow canker growth resistance trait, provides direct context for this research. Before addressing the research direction of this thesis, the above topics are reviewed in sections 1.1 through Biology of C. ribicola White pine blister rust is an introduced disease of 5-needle pines caused by Cronartium ribicola, a heteroecious, macrocyclic rust fungus that likely evolved during the Miocene epoch in central Eurasia near the Ural mountains (Hummer 000). North America contains 9 of the world s nearly 0 species of white pines, followed by Asia, Europe and the Mediterranean region (Maloy 1997). Shrubs in the Grossulariaceae family, such as currants and gooseberries, are the alternate host required for C. ribicola to complete its 5-stage life cycle. Pedicularis spp. and Castilleja spp. were also recently identified as secondary hosts for C. ribicola (McDonald et al. 006). Blister rust causes branch and stem-cankers on white pines that are typically diamond-shaped and delineated by orange margins (Fig 1). Although blister rust mainly affects the phloem of branch and stems, infection by C. ribicola is

17 3 initiated in the needles. Spores land on needles, germinate through stomata, and hyphae subsequently grow through needle tissue and into phloem. The pathogen produces 3 spore types (urediniospores, teliospores, and basidiospores) on Ribes spp., and spore types (pycniospores and aeciospores) on Pinus spp. Relative spore hardiness and spore travel distances vary (Table 1). All 5-needle pines are susceptible to infection by C. ribicola, and only 1 of the 8 native species of 5- needle pines in the U.S. has not yet been infected (Table ). Fig 1. Blister rust stem canker with typical morphological characteristics

18 4 Table 1. The 5 spore types of C. ribicola vary in relative hardiness and traveling distance. Spore Type Host Ploidy Level Spore Description Pycniospores Pinus sp. N Aeciospores Pinus sp. N Urediniospores Ribes sp. N Teliospores Ribes sp. N Basidiospores Ribes sp. N Spores serve sexual function when insects vector genetic material from one pycnial droplet to another within the same canker and to other local cankers. Powdery yellow, football-shaped aeciospores (1X30 micrometers in size) are windborne up to 500 kilometers to the Ribes sp. host. Thick-walled and relatively resistant to dessication. Produced on ventral side of Ribes sp. leaves. Spores are relatively fragile and increase local inoculum as they repeatedly reinfect nearby Ribes sp. leaves throughout the summer Germinate in place in hair-like columns on ventral side of Ribes sp. leaves. Requires high relative humidity to stay viable Thin-walled and relatively fragile. High humidity and cool temperatures required for travel to local white pine hosts

19 5 Table. The causal pathogen of white pine blister rust, C. ribicola, has affected 8 of the 9 U.S. native species of white pine. Common Name Latin Name Status Eastern white pine Pinus strobus L. Infected Western white pine Pinus monticola Dougl. Infected Sugar pine Pinus lambertiana Dougl. Infected Limber pine Pinus flexilis James Infected Bristlecone pine Pinus aristata Englem. Infected Whitebark pine Pinus albicaulis Dougl. Infected Southwestern white pine Pinus strobiformis Englem. Infected Foxtail pine Pinus balfouriana Grev. & Balf. Infected Great Basin bristlecone pine P. longaeva D.K. Bailey Not infected Pycniospores Pycniospores appear on Pinus spp. either during the same or second season following fungal establishment in the phloem (Lachmund 1933). During the summer, pycnia in Pinus spp. phloem produce pear-shaped haploid pycniospores (.8 µm) at the mycelial margin of the canker (Hirt 1964). Pycniospores provide an early visual confirmation of active blister rust disease on white pine: rust colored, spore-filled droplets exuding from blister-like bumps on an infected stem or branch. The fungus over-winters in the Pinus spp. host, until dome-shaped aecia begin to form under the bark during the following spring Aeciospores Aeciospores are the last of spore stages appearing on the Pinus spp. host, typically -3 seasons after initial infection. Aecia usually develop in the same area as

20 6 the previous pycnia. Each aecium expands with multiplying aeciospores until the overlying bark ruptures, exposing masses of powdery yellow, football-shaped, binucleate aeciospores (1X30 µm) in late spring/early summer (McDonald and Hoff 001). Aecial masses are exposed in an annual progression of bark desiccation that progress around the stem or branch. Aeciospores are windborne up to 500 kilometers (Mielke 1943) to the Ribes spp. host where they germinate through abaxial stomata Urediniospores Urediniospores are the first of 3 spore stages on the Ribes spp. host. After aeciospores land on Ribes spp. leaves, they germinate into orange horseshoe-shaped sori (about 0.5 mm diameter) called uredinia. Uredinia produce multiple generations of orange, binucleate urediniospores that repeatedly reinfect the alternate host (Ribes spp.) throughout the summer (Colley 1918). Amplification of the fungus on Ribes spp. through the uredinial infection cycle can result in increased inoculum levels of the basidiospore stage, which is infective on the primary host, Pinus spp Teliospores Night time temperature decreases in late summer stimulate the formation of telia that produce oval, binucleate teliospores (about 10X40 µm) packed in hair-like telial columns (up to 1500 µm long) on the abaxial side of Ribes spp. leaves. Teliospores, the second of 3 spore types on Ribes spp., stay packed in their telial

21 7 columns until germination is triggered by 6 to 8 hours of near 100 percent relative humidity in the fall. (McDonald and Hoff 001) Basidiospores Four spherical, binucleate basidiospores (about 5 to 10 micrometers in diameter) are produced by each germinated teliospore, and each telial column produces about 6,000 basidiospores (McDonald and Hoff, 001). Basidiospores, the third spore stage on Ribes spp. leaves, are windborne to nearby Pinus spp. needles, where they germinate through stomata and colonize needle tissue in the fall. Hyphae grow through the needle tissues and into branch phloem. 1. Spread of White Pine Blister Rust 1..1 North American Introduction of C. ribicola The earliest North American introduction of the rust fungus C. ribicola was reported in 1897 in York County, Maine on European black currants (Ribes nigrum) imported from England (Posey and Ford 194). Although white pine blister rust was first found on white pines in a tree nursery near Philadelphia, PA in 1905 (Pierce 1917), eastern U.S. infection centers were already established as millions of diseased white pine seedlings from European nurseries were planted in more than 00 northeastern locations between 1900 and 1910 (Filler 1933). The Plant Quarantine Act of 191 controlled the entry and movement of nursery stock (Maloy 1997), but was

22 8 too late to keep the disease out of western North America. Although early scouting found western forests blister rust-free until 191 (when blister rust was first found in southwestern British Columbia and northwestern Washington), later investigations revealed the original western introduction of the pathogen occurred in 1910 at Point Grey near Vancouver, B.C. on infected white pine seedlings imported from French nurseries: Pierre Sebire and Son Ussy (Lachmund 196). It is generally accepted that the 1910 importation of infected seedlings at Point Grey is the first introduction of C. ribicola in western North America. A short time before the discovery of the 1910 introduction, Perley Spaulding prophetically wrote this caution about blister rust traveling to the western U.S. on infected seedlings: A single diseased shipment may undo all attempts to restrict it to the eastern forests (Spaulding 19). From the 1910 introduction, blister rust disease eventually spread throughout the entire range of 5-needle pines in the West. 1.. Intensification of White Pine Blister Rust Intensification of the disease occurs when multiple generations of urediniospores on Ribes spp. reinfect and spread to neighboring plants. Pinus spp. infection occurs when conditions are favorable for basidiospore infection. Greater numbers of uredinia leads to greater numbers of resulting Basidiospores. Higher amplification of spores on Ribes spp. leads to greater disease pressure on pines. A single founding aeciospore, given the right environmental conditions, can therefore begin a new blister rust colony (Kinloch 003). Historically, the leading edge of blister rust appeared first on Ribes

23 9 spp. plants, sometimes hundreds of km away from the nearest infected white pine (Mielke 1943). This phenomenon initially confused blister rust surveyors who found basidiospores on Ribes spp. leaves sometimes hundreds of kilometers away from any infected pine trees. Further adding to the confusion of understanding the mode of blister rust spread, infected Ribes spp. leaves were abscised in the fall, many months before any sign of infection appeared in the pine host the following spring Wave Years Strong winds, low precipitation, and moderate temperatures in the spring can combine to provide a favorable climate for the long-distance dissemination of aeciospores (Mielke 1943). Regional weather data linked with historical observations show that infection of Ribes spp. is most intensified by cool, wet cycles during the summer (McDonald et al. 1981). Furthermore, teliospore germination, basidiospore formation, and subsequent Pinus spp. infection also require high relative humidity coupled with cool temperatures (Spaulding 19). When favorable environmental conditions are combined, blister rust spreads into new white pine territory in wave-like expansions called wave years Western Spread of C. ribicola In 191, 10 years after its western introduction on white pine seedlings, the discovery of the fungus on planted European black currant inaugurated U.S. and Canadian scouting efforts to determine the range of the disease (Meilke 1943,

24 10 Pennington 195). U.S. and Canadian inspections found the rust at many sites located in the coastal pine region of British Columbia and northwest Washington (McDonald and Hoff 001, Lachmund 196). Subsequently, in order to track the pathogen s spread on wild Ribes spp., forest plots were established in Washington and British Columbia. While the Canadian scouting program ended in British Columbia in 1930, the U.S. scouting program continued to track the spread of blister rust on wild Ribes spp. in the West for several more decades (Meilke 1943). The single 1910 introduction of C. ribicola in British Columbia afforded blister rust workers a better opportunity to track blister rust spread than did the many dispersed introductions in the East. It is generally accepted that the first post-introduction wave year occurred in 1913 for reasons: weather records show that 1913 was a favorable year for blister rust development, and cankers with 1913 inception were as far as 09 km northwest, 177 km north and 113 km east of Pt. Grey (Lachmund 196). For similar reasons, 1917, 191, and 193 (when blister rust invaded the Inland Empire) are also considered wave years. Several cool, moist periods in 195 prolonged aeciospore production, and the rust was distributed on Ribes spp. over a greater range than ever before: Ribes spp. infection was found 748 km south, and 644 km east of Pt. Grey (Mielke 1943). By 198, the rust was established in northern Idaho, western Montana, and the Oregon Cascades. Blister rust reached northern California white pines in 1930, but the very dry summer of 1931 reduced aeciospore and basidiospore production, slowing both the spread and intensification of the disease. Blister rust was found east of the Continental Divide in 1937, which was the most significant wave year since 197.

25 11 During 1937, favorable weather and wind conditions helped spread the disease on Ribes spp. in southwest Oregon and northern California. High aeciospore production and strong northerly winds in 1938 carried blister rust to California s Sierra Nevada range (Mielke 1943). As blister rust continued to spread throughout the West, alarming reports of its invasion into high elevation white pine ecosystems emerged. Blister rust reached foxtail pine in northern California in 194 (Miller 1968), and southern Idaho limber pine in 1945 (Krebill 1964). Threateningly close to California s southern Sierra Nevada populations of foxtail and limber pines, blister rust infected sugar pine in the southern end of the Sierra Nevada in 1961 (Kinloch and Dulitz 1990). Infection was found on isolated patches of southwestern white pines growing high in the Sacramento Mountains of New Mexico in 1990 (Hawksworth 1990). Blister rust, as of 006, had not been found in Arizona or mainland Mexico (Geils 006, unpublished data). Nor, as of 003, had it been found in the high elevation stands of bristlecone pine in the White Mountains of California (Kinloch 003). 1.3 Economic and Ecological Impacts of Blister Rust Western white pine was historically the most important timber species harvested in the Rocky Mountain area (Benson 1967). Early harvesting practices, blister rust disease, and the decline in the wood match industry combined to bring western white pine down to a much lower level of economic significance. The combined value of the 3 most economically important species of white pine timber (P.

26 1 lambertiana, P. strobus, P. monticola) was estimated to be over a billion dollars in 1919, and by 199, the Inland Empire was producing over 43% of U.S. white pine (Maloy 1997). In 1945, Idaho produced 0,31 million board feet of western white pine (Steer 1948). By 1981, net growing stock volume of western white pine was 1,33 million board feet, and comprised only 4.3% of total net growing stock volume of all timberlands in Idaho (Benson et al. 1987). By 1996, only 5% of the Inland Empire western white pine growing region was re-planted with western white pine (Fins et al. 001). The most apparent ecological impact of blister rust in historically western white pine regions is the conversion of western white pines to more shade-tolerant, droughtintolerant species such as western hemlock (Tsuga heterophylla (Raf.) Sarg.), Douglas-fir (Pseudotsuga menziesii (Mirbel) Franco), and grand fir (Abies grandis Dougl. ex D. Don). These species become stressed during periodic droughts and are highly susceptible to bark beetles and native root and foliage pathogens (Hagle et al. 1989) 1.4 Blister Rust Control Through Ribes spp. Eradication C. ribicola is heteroecious, and eliminating Ribes spp. plants from white pine ecosystems can theoretically break the pathogen s life cycle. This theory proved reasonably successful when put into practice in the eastern United States. Blister rust decreased in the east as a result of laws like Quarantine Number 6 (enacted April 1, 1917), that not only prevented people from planting ornamental Ribes spp. bushes, but

27 13 also required destruction of all Ribes spp. plantations (Maloy 1997). Inspired by success in the east, the USDA Bureau of Plant Industry initiated a Ribes spp. eradication program in the west and by 197, all known Ribes spp. plantations were destroyed in Washington, Idaho, Montana, and most of California. Canada, however, chose not to eliminate their R. nigrum plantations, and those plantations played an important role in the spread of the rust to white pines in the interior white pine belt (Meilke 1943). During the eradication program, blister rust control crews covered millions of acres of western white pine forests, hand pulling and spraying upland and lowland Ribes spp. plants. Ribes spp. plants in Western white pine forests proved too ubiquitous to remove entirely from the ecosystem. Although the program failed to eradicate the secondary host from western forests, the Blister Rust Control program provided many people with an income and a sense of purpose during America s Great Depression. 1.5 Resistance Breeding Program Active research on white pine blister rust resistance in the western U.S. dates from Researchers began the breeding program in the Intermountain West by identifying candidate resistant trees in the wild and collecting their cones and pollen for controlled crosses (Bingham 1983). By 195, Bingham and others proved that resistance of the wild survivors was inherited. Three elevation zones, low (< 1067m), mid (1067m 150m), and high (> 150m) were recognized, and growing and planting sites were established. Researchers estimated these 3 zones would cover the

28 14 total 810,000 hectares of planting sites in the Inland Empire (Bingham 1983). Thousands of trials with wild parent trees were performed in the 1960s, and by the early 1970s, the first generation from initial controlled crosses (F1 stock) was ready to grow in Inland Empire seed orchards. From 1970 to 1985, seed orchards produced about 590 kg of F1 and F (F1 progeny) seeds, but this still was not enough to meet the demand from the timber industry (Goddard et al. 1985). Significant resistance gains were realized in 1973 when researchers reported that F seedlings were 66% uninfected ½ years after inoculation with C. ribicola (Hoff et al. 1973). This was a 33% increase in resistance from F1 seedling results. By the early 1980s, F stock had consistently resisted infection under intense rust exposure, and researchers estimated long-range survival of F stock under natural rust exposure would exceed 65% (Bingham 1983). Subsequent surveys of Idaho F field resistance found average infection rates were much higher (Schwandt and Ferguson 00). Environmental effects were found to have a significant impact on resistance in Idaho F stock in British Columbia (Hunt 004), proving that durability of resistance responses depends on environment as well as mode of action. Both rust hazard and the evolutionary potential of C. ribicola, therefore, are important considerations when deploying resistant stock (Sniezko and Kegley 00). More work is needed to fully evaluate field resistance of western white pine in the Inland Empire, and geneticists are looking for more resistant phenotypes on which to focus breeding efforts. Slow (or abnormal) canker growth is one area to focus field evaluations on. Field assessments of abnormal

29 15 canker growth can supply breeding programs with valuable information about one possible phenotypic resistance response. 1.6 Slow Canker Growth Resistance Trait Slow canker growth, although variously described (Hoff and McDonald 1980, Hunt 1997), is believed to be a heritable, polygenic resistance trait that allows greater survivability of infected trees (Hunt 1997). Unlike major gene resistance mechanisms (such as needle spot reactions), slow canker growth avoids putting strong selection pressure on the fungus (Hoff 1984). Cankered, living trees allow C. ribicola to survive and complete its life cycle, thereby decreasing pressure to develop new races that could neutralize the resistance effect (Bingham et al. 1971). As resistance terminology evolved, slow canker growth and tolerance resistance types were combined into one category generally thought of as higher survival of trees with stem symptoms (Sniezko and Kegley 00). Research that followed over 3,000 stem-cankered seedlings from original progeny tests planted in Idaho found 58 trees still surviving with stem-cankers 30 years later. All of those stem-cankers were abnormal in appearance (Hoff 1984). Researchers believe that alive and cankered is the last of several resistance mechanisms to be expressed, and should be a priority mechanism for selection (Hoff 1984).

30 16. Thesis Research Direction Researchers are noticing live and stem-cankered trees in pole-sized F western white pine, and realize that the frequency of abnormal stem cankers is increasing. Surveys are presently underway to collect canker frequency and tree mortality data in permanent F western white pine plots (Schwandt and Ferguson 00). Little is known about the relationship between apparently abnormal stem-canker morphology and circumferential tissue damage over time. In 1967, eastern white pine stem-cankers were visually classified into 3 distinct types after researchers followed their growth, development, and sporulation for 5 years (Phelps and Weber 1968). Phelps and Weber used the same physical criterion to type cankers, and their 3 different types are identical to the 3 different classes presented in this research. Based on 400 bole cankers on eastern white pine ranging in age from 8 to 31 years, they concluded that canker growth slows down with increasing age of host and canker, and the slowestgrowing canker type appeared to be inactivated. No known follow-up research was conducted on Phelps and Weber s canker types on eastern white pine (Ostry, M.E. personal communication February, 007). Even though cankers on seedlings in breeding orchards are carefully observed and measured, more work is needed to assess field performance of F stock under natural conditions (Sniezko and Kegley 00). Forest managers, however, typically do not have time or money to follow stem canker development over time. Furthermore, most forest managers expect all stem-cankered white pines to be killed within several years. Methods in this research outline one way forest managers can quickly classify stem-

31 17 cankers on pole-sized F western white pine in order to make assumptions about F field performance. The motivation for this study is to show that stem-cankers on pole-sized Idaho F western white pine trees can be separated into severity classes based on external canker characteristics. Cankers in the least-severe Class, for instance, may indicate that their host will survive longer than expected; possibly long enough to meet stand management objectives. Foresters more familiar with F field performance will be better equipped to predict stand volume, write pruning and thinning prescriptions, and guide future stand composition and structure. The nature of this project is observational, and project goals are strongly oriented to extension and applied forest management. Inland Empire forest researchers observed scattered F western white pine with abnormal-looking stem-cankers growing in the same plantation as trees with normallooking stem-cankers. Atypical stem cankers departed from normal diamond-shaped bark discolorations, and ranged in shape from round depressions to long, vertical stem crevices. After further observation, it seemed that cankers could be classified into 3 general canker classes, and we hypothesized that they are linked to canker age as well as canker severity (amount of pathogen-caused tissue damage over time). The 3 classes of cankers were Class I (high severity), Class II (medium severity), and Class III (low severity). Some cankers appear to grow rapidly with no apparent defense response from the tree; we hypothesize cankers in this class (Class I) damage the most host tissue over time. Other cankers seem to grow slower and are bounded by marginal

32 18 swelling that appears to be a tree defense response; we hypothesize cankers in this class (Class II) are older than Class I cankers and are damaging relatively less host tissue over time. Finally, Class III cankers appear completely inactivated by their host, existing only as vertical stem depressions ; we hypothesize cankers in this third class are relatively the oldest cankers and damage the least amount of host tissue over time. In short, the interactions of the fungus and its host lead to variation in progress of tissue damage and to potential host longevity. Although the statistical scope of inference of this research is limited to pole-sized Idaho F western white pine trees planted in the 3 northern Idaho plantations described herein, we have no reason to expect that abnormal stem cankers in this study would be much different than abnormal cankers in other F plantations. The 3 main goals of this research on white pine blister rust stem cankers in northern Idaho are: (1) determine if this canker classification system can give reproducible results, () determine if cankers in different classes affect relative tree growth differently, and (3) determine if tree circumference is related to amount of pathogen-caused first-year tissue damage.

33 19 3. Materials and Methods 3.1 Canker Class Definition Class I Cankers For inclusion in Class I, many cankers had bright, obvious, mycelial margins that formed the typical diamond shape around infection entry with no swelling of margins or obvious hypersensitive reactions (Fig. ). Fig. Class I canker with typical characteristics: no marginal swelling or obvious tree reaction, orange mycelial margin forms diamond shape around branch of infection entry

34 0 Occasionally, some cankers had no clear orange margins or callous formation but obviously appeared to be active with large amounts of fresh resin, without apparent resistance response from the tree. These were also typed as Class I cankers (Fig. 3) Fig 3. Class I canker without orange margin, but obviously active around entry branch with no marginal swelling or apparent tree reaction

35 Class II Cankers Cankers placed in Class II were bounded by tree tissue swelling and necrotic lesions, were usually broader horizontally than long, orange margins were less obvious or absent, and resin was usually recent. (Fig 4). Fig 4. Class II canker with obvious swelling around entry branch and recent resin

36 3.1.3 Class III Cankers The third class contained cankers with no recent sign of resin, no orange margins, and marginal swelling that created a narrow depression, and no other signs of fungal activity (Fig 5). Fig. 5. Class III canker that created narrow depression in stem with no signs of fungal activity

37 3 3. Site Selection and Description Collaboration with land managers in the Idaho Department of Lands, the St. Joe National Forest, Potlatch Corporation, and the University of Idaho provided a list of approximately 500 rust-resistant western white pine plantations that met the following criteria: 1) within 1km from forest roads, ) planted prior to 1994, 3) well-stocked with surviving western white pine, 4) within 30km of the USDA Forest Service Forest Health Protection laboratory in Coeur D Alene, Idaho, and 6) unthinned. From this list, 3 rust-resistant plantations were selected (Table 3). 3.3 Canker Selection In order to systematically search for the 3 canker categories, flagged transects were installed across each plantation every chains (40 m). Approximately 0 cankers in each Class were flagged in each stand. Aluminum nails were used to attach numbered tags next to selected bole cankers. 3.4 Exterior Canker Measurements After canker selection, the following 7 measurements were made on the stemcankers while trees were still standing: 1. Canker height (nearest cm.) from groundline to entry point of infection. Canker length (nearest cm) from top to bottom of the canker on the outside of stem 3. Canker width (nearest cm), across widest part of the canker on the outside of stem 4. Type of canker (I, II, III), as previously described 5. Diameter of tree at breast height (DBH) (nearest cm) 6. Diameter of tree at center of canker (nearest cm) 7. Aspect of canker (degrees)

38 Table 3. Descriptions of 3 plantations from which cankers were harvested. Ownership Hectares Year Planted Aspect (AZ) Average Slope (%) Elevation (m) Estimated Infection % UTM Coordinates at Site Center Easting Northing Associated Tree Species Clarkia Kingston Rocky Point US Forest Service Idaho Department of Lands University of Idaho t t t UTM UTM UTM ponderosa pine Douglas-fir ponderosa pine Douglas-fir western larch ponderosa pine Douglas-fir 4

39 5 3.5 Canker Harvest Tagged trees were cut with a chainsaw, and tree height (nearest.1 m) was measured once the tree was on the ground. Cankered sections of the trees were cut from the stem. We used u-shaped fencing staples to attach pieces of rope to the ends of the logs in order to drag the cankered sections out of the woods. The cankered logs, averaging about 18 cm in diameter and 1 m long, were transported to the Forest Science Laboratory at Oregon State University. 3.6 Canker Cross-Sectioning I used a 7.5 horsepower band saw with a 91cm throat to make cross-sectional cuts on each cankered log. The first cut bisected the cankers, exposing annual growth rings and revealing canker history. The second cut created a 5cm thick cross-sectional slice (stem section) (Fig 6). In order to find the first year of stem xylem disruption and expose subsequent canker development, the first cut was made at the estimated fungal entry point into the stem. C. ribicola enters stem phloem primarily through infected branches but less commonly through stem needles. A dead branch stub protruded from the center of most cankers in this study, pinpointing location of pathogen entry. In the few cases where location of infection entry was unclear, the first cross-sectional cut was made at the center of the canker. Exploratory slices through cankers without a branch stub indicating a clear entry point confirmed that cutting through the center was usually the

40 6 best approach to finding the oldest damaged annual ring. The 5cm stem section provided the basis for all tissue damage measurements. Fig 6. Stem sections cut from cankers on a 7.5 hp bandsaw were 5 cm thick 3.7 Photographing Cross Sections Using a tripod, I took a digital picture of each 5 cm stem section from a fixed height of.5m. To provide scale and calibration, each slice was photographed on a cm by cm grid (Fig 7). Photographs were taken immediately after sawing, because resin usually flowed over annual rings immediately after cutting, making healthy xylem indistinguishable from damaged xylem.

41 Fig 7. Digital photograph of a 5 cm thick stem section from a standard fixed height. Stem section was cross-sectioned at the point of infection entry and photographed on a by cm grid for scale. 7

42 8 3.8 Analyzing Cross Sections Photographs were saved as.jpg files and imported into a Windows (Win3) application called Reconstruct (Fiala J.C. 005). Reconstruct facilitated precise linear and area measurements of the photographed stem sections. The following 5 measurements were made on each image (Fig 8): 1. Annual ring circumferential length (cm) for each year beginning at pith. Annual ring area (cm ) for each year beginning at pith 3. Tree age at canker initiation (yr) 4. Canker age (yr) for all years of stem tissue damage 5. Length of circumferential tissue damage (cm) for all canker years Circumferential Infection Damage Annual Ring Area Annual Ring Length Fig 8. Digital photograph of cankered stem section showing traced annual ring length, ring area, and circumferential tissue damage

43 Measuring Annual Ring Circumferential Length and Area I traced undamaged annual ring length and annual ring cross-sectional area for every year individually on each cross-section image (Fig 8). Annual ring length provided the basis for measuring circumferential proportion of healthy tissue to damaged tissue for every year since canker inception. Annual ring length measurements can also be converted to tree diameter at canker height for any given year Measuring Circumferential Tissue Damage Magnification of each stem section image within the program Reconstruct allowed precise measurement of canker-caused circumferential tissue damage for each canker year. Tissue necrosis follows along late-wood ring patterns. Visible tissue damage was used as the measurement of canker progression and is defined as the length (cm) of annual ring that appeared black, dead, non-functional, or missing. Continuation of annual rings is halted wherever the pathogen kills cambial cells. In cases where an annual ring stopped, I continued to trace over the gap where the annual ring otherwise would have been formed. Tree circumferences (for percent girdling calculations) were therefore established by projecting the missing part of annual rings. Percent girdling is defined as the length of damaged annual ring divided by the total annual ring length x 100.

44 Statistical Analyses Since some Class I cankers with bright orange margins had not yet disrupted or killed any host tissue, there was nothing on which to base canker progression, and approximately 10 of those were discarded from the analysis. I assumed these cankers had yet to produce damaging pycniospores or aeciospores. Coalesced cankers were also impossible to measure accurately and were discarded. No measurements on cankers other than the ones of interest were taken. All analyzed cankers occurred as single bole infections below 1.5 m. Finally, some cankers simply had too much resin to distinguish annual rings or tissue damage. Altogether, approximately 50 harvested cankers were discarded before reaching any stage of analyses Relationship Between Percent Girdle and Canker Class A randomized block, repeated measures, observational study was conducted to determine whether percent girdle by stem-cankers was related to canker age, and whether this relationship was the same for the 3 canker categories. Canker-caused circumferential tissue damage (response variable) was measured annually on each tree (sampling unit) from canker inception (canker age 1) to canker harvest, and no randomization of sample selection, treatment, or sampling unit occurred. Since tissue damage was repeatedly measured on the same trees over time, significant correlation of response variable measurements exist. Any two measurements of tissue damage on the same tree, no matter how separated in time they occur, are similarly influenced by the shared condition of the tree (they co-vary). Additionally, any two measurements