ABSTRACT

Atmospheric nitrogen (N) fixation by Alnus tenuifoliacan account for up to 70%, of the N accumulatedduring vegetation development along river flood-plains in interior Alaska. We assessed disease inci-dence and related mortality of a recent outbreakor fungal stem cankers on A. tenuifolia across threeregions in Alaska during the 2005 growing season, anddetermined the impacts on N-fixation rates, nodulebiomass, and stand-level N-fixarion inputs. Thehighest percentage of ramets colonized or dead withcanker was found on Tanana River plots, suggestingthe epidemic is most severe in the Fairbanks region. Apositive relationship between % basal area loss tocanker and % canopy loss provides a simple means forassessing stand-level mortality associated with diseasein the field. Although specific N-fixation (SNF) rateswere not influenced by canker disease incidenceof individual genets, live nodule biomass beneathalder canopies was inversely correlated with the per-

centage of ramets dead or with main ramet canker.Variations in SNF and live nodule biomass translatedto differences in N-fixation inputs, which ranged from22 to 107 kg N ha-1 y-1 across study regions. Nodulebiomass was reduced by incidence of canker diseaseand related mortality an average of 24% across allsites, which translates to N input reductions of 8, 16,and 33 kg N ha-1 y-1 for the three regions, respec-tively. During the 2008 growing season, we resur-veyed the Tanana River plots and found that of theramets larger than 4-cm diameter baving main rametcanker in 2005, 74% are now dead; and for thosewithout main ramet canker in 2005, 25% havedeveloped main ramet canker, and 8% are dead. Thus,it is likely that N-fixation inputs have declined furtherbelow what we estimated for 2005.

Key words: Alaska; alder; canker; disease; nitro-gen cycling; nitrogen fixation; succession.

INTRODUCTION

Plant pathogens and invertebrate herbivoresinfluence important pathways for carbon (C) andnitrogen (N) cycling in terrestrial ecosystems. Thisincludes direct effects on plant growth, mortality,foliage chemistry, and canopy structure (Hansenand Goheen 2000; Tainter and others 2000; Erickson

Disease-Mediated Declines inN-Fixation Inputs by Alnus tenuifolia

to Early-Successional Floodplainsin Interior and South-Central Alaska

Roger W. Ruess.1* Jack M. Mcf'arland,1 Lori M. Trummer,2 andJennifer K. Rohrs-Richey1

michroa croceas and invasive (Eriocampa ovatas aldersawflies, regional climate warming. moisture stress,and branch damage from frost, heavy snowloads.or river ice. South-central and interior Alaskaexperienced the hottest summer on record during2004 (http://www.wrcc.dri.edu/summary/clmis

mak.hrml).which reduced alder growth (Nossov2008),and may have created stressful host condi-

tions and a favorable environment for rapid diseasedevelopment. In addition to stressful abiotic agents.several other opportunistic canker fungi have beenfound in association with diffuse cankers on A.tenuifolia (USDA 2008), and it is likely that anumber of biotic agents are contributing to diseaseincidence or accelerating the rate of dieback aridmortality. It is possible that canker diseases may benatural components of the long-term populationbiology of the host, evidenced by the spread ofdisease coincident with the recruitment of early-successional alder stands along the Tanana Riverbasin over the past 20 years (Nossov 2008). How-ever, the widespread increased incidence of cankeron A. tenuifolia throughout both Alaska and theRocky Mountain states (Worrall and Adams 2006;Worrall and others 2008; Worrall 2009) suggeststhat the epidemiology of the disease is morecomplex.

Given that up to 70% of the N accumulatedduring forest succession along Alaskan floodplainscan be derived through atmospheric fixation by A.tenuifolia (Van Cleve and others 1971; Uliassi andRuess 2002), we hypothesized that high diseaseincidence and associated mortality of alder wouldhave significant effects on ecosystem N inputs de-rived from atmospheric fixation. Differences in N-fixation rates of individual A. tenuifolia plants at theregional scale are expected to be driven by varia-tion in plant growth and N demand. and influencedby site conditions such as light, soil moisture, andsoil nutrient availability, notably phosphorus (P)(Uliassi and Ruess 2002). Disease - related declinesin N-fixation inputs at the stand-level could bemanifested through reductions in nodule biomassresulting from ramet mortality, and/or decreases inN-fixation rate per unit nodule due to physiologicaldown-regulation of nitrogenase activity (Ruess andothers 2006). Such coupled effects would ulti-mately influence soil N accumulation, forest standdevelopment, and forage quality for higher trophiclevels (Butler and Kielland 2008).

This study had two objectives. The first was toassess the incidence of canker disease and disease-related mortality of A. tenuifolia across Alaska bysurveying a set of sites throughout interior andsouth-central Alaska. The second objective was to

and others 2004; Stadler and others 2006) andlitter composition and decomposability (Waringand others 1986; Cobb and others 2006). Plantpathogens can also have strong indirect effects onnutrient cycling through effects on soil microcli-mate (Jenkins and others 1999), rates of herbivoryand trophic structure (Faeth and Hammon 1997;Tingley and others 2002) throughfall chemistry(Lightfoot and Whitford 1990; Reynolds and Hun-ter: 2001), and the quality and quantity of allo-chthonous inputs to aquatic systems (Snyder andothers 2002). Pathogen and invertebrate herbivoreimpacts on foundation species. often have thegreatest consequences for changes in nutrient cy-cling processes because such hosts, by definition,play a prominent role in ecosystem structure andfunctioning (Ellison and others 2005). Althoughecosystems are expected to be resilient to outbreaksof native herbivores and pathogens, invasive her-bivores and pathogens are often associated withpermanent shifts in ecosystem structure and func-tion, including significant changes in nutrient cy-cling dynamics (Castello and others 1995; Ellisonand others 2005; Jenkins and others 2007. How-ever, climate change may alter the interactionsbetween plants and their native pathogens andinvertebrate herbivores (Malmstrom and Raffa2000); Chakraborty and Datta); 2003; Scherm 2004;Mulder and others 2008), resulting in regime shiftsthat permanently alter nutrient cycling dynamics.This may occur through direct effects of climatewarming on a whole suite of factors. These includechanges to insect or pathogen life cycles (Berg andothers 2006), changes in plant vulnerability toinfestation by altering the interactions betweenpathogens and invertebrate herbivores (Mulderand others 2008), or by affecting interactions be-tween plant parasites and other disturbance re-gimes affected by climate change, such as fire(Parker and others 2006; Allen 2007).

Observations by the Alaska USDA Forest ServiceForest Health Protection staff of widespread diebackand mortality of thin-leaf alder (Alnus incana ssp.tenuifolia; hereafter, Alnus tenuifolias throughoutsouth-central and interior Alaska has recently fo-cused attention on the extent, potential causes andconsequences for this mortality (USDA 2008).Dieback and mortality throughout Alaska has beenmost frequently associated with a fungal cankerpathogen (putatively Valsa melanodiscus, anamorphCytospora umbrina) which causes diffuse, longitudi-nally expanding cankers that eventually girdle thestem. There are many factors that may potentiallyinteract to increase predisposition or severity ofdisease, including defoliation by the native (He-

stands 15-25 years of age dominated by densestands of A. tenuifolia. Scattered balsam poplar(Populus balsamifera L.) saplings are recruiting intothese stands, and a number of herbaceous speciesconstitute a ground cover where canopy gaps per-mit light penetration to the ground surface.

SoilsTen soil cores (2.54 cm diameter x 15 cm depth)were sampled from each of the nine plots across thethree regions. Soils were sampled from beneatheach shrub used for N-fixation determinations (seebelow) after removing the surface litter layer. Soilswere stored in 2 mil plastic bags on ice for up to3 days and processed within a day of being broughtto the laboratory in Fairbanks. The top 15 cm ofeach core was homogenized by removing largeroots and organic debris, passed through a 2-mmsieve, dried to constant weight at 45°C, and groundin a steel ball-mill. Soil pH was determined on a 2:1(water/soil) extract, and total C and N were deter-mined on a LECO CNS 2000 (Leco Corporation, St.Joseph, Michigan, USA). Total soil P was deter-mined colorimetrically on an autoanalyzer follow-ing perchloric acid digestion. Available P wasdetermined by extracting 109 soil in 100 mldeionized water containing a fine-mesh nylon bagcontaining 4 g Bio-Rad AG l-X8, 0.50 mol 1-1

NaHC03-charged anion exchange resin, followedby 0.50 mol-1 HCL resin bag exchange andautoanalyzer analysis (Lajtha and others 1999).

Vegetation and Canker SurveyA. tenuifolia genets grow as single or multiple mainstems (ramets) arising at or just above the soilsurface. Within each plot, the diameters of all

determine the impact of disease incidence andrelated mortality on Nvfixation rates and stand-levelN inputs derived from N fixation by A. tenulfolia.

MATERIALS AND METHODS

Study SitesDuring the slimmer of 2005, permanent study plotswere established in three regions spanning a lati-tudinal transect from the Kenai Peninsula to justsouth of Fairbanks, Alaska. River drainages whereA. tenuifolia was a dominant or co-dominant specieswere selected for study within each region. Theseincluded Quartz Creek (Kenai region), Eagle River(Anchorage region), and Tanana River (Fairbanksregion) (Table 1). At eachlocation, three 20 x20 m2 plots were established within a 2-3 kmstretch of the river.

Plots along Quartz Creek had the lowest stemdensity and basal area of A. tenuifolia, and weresituated on older stable terraces above the river asevidenced by the maturity of scattered white (Piceaqlauca (Moench) Voss) and Lutz (Picea x lutzii)spruce, and Alaska paper birch (Betula neoalaskaSarg.). Willows (Salix sp.) and rose (Rosa acicularisLindl.) were scattered amongst the plots, withhorsetails (Equisetum sp.) and bluejoint reedgrass(Calamaqrostis canadensis) forming a dense groundcover in partially shaded and open stands, respec-tively. Within the Eagle River drainage, A. tenuifoliagrows in dense stands along low-lying terracesthat are flooded during spring snowmelt. Standsselected for study were nearly pure alder, with C.canadensis forming a dense ground cover in openstands. and white spruce recruiting in more closedstands. Along the Tanana River south-west ofFairbanks. plots were located in early-successional

percent enrichment (APE) of 15N2 at time zero (T0).These samples were stored in 10 ml exetainers(Labco. High Wycombe, Buckingharnshire) beforeanalysis using a dual-inlet isotope ratio mass spec-trometer (PDZ Europa Scientific Instruments.Crewe, Chesire). After the removal of the incuba-tion atmosphere sample, the syringe containing theroot nodules was immediately returned to the for-est floor for 10 min. Nodules were then removedfrom the syringe and immediately frozen in liquidN2. Soil moisture (TDR, CS620 Hydroserise with10 cm probes, Campbell Scientific, Logan, Utah)and temperature (digital thermometer, 7 cm depth)were measured at live locations beneath the can-opy of each genet.

In the laboratory, nodules were thoroughlyrinsed through a fine sieve of all adhering soil andorganic material, dried for 48 h at 60°C, andground using a Wig-L-Bug ball-mill (Rellex Ana-lytical, Ridgewood, New Jersey) in preparation formass spectrometry analysis. A nodule sample notincubated with 15N2 was used as a control for thedetermination of APE for each nodule sampleaccording to the following equation:

where both l5N content measures are in at. %. Bycombining APE with total nodule N content, divid-ing by incubation time, and correcting for thecomposition of the initial incubation atmosphere asdetermined by mass spectrometry, we calculated thespecific N-fixation activity of the nodule samples(SNF = umol N assimilated gDWT nodule 1 h 1) asfollows:

where %, Nnodule is the mass percent N content ofthe enriched nodule sample and % 15Natmoshere the atom percent 15N content of the incubationatmosphere at the beginning of the assay.

The incidence of disease or disease-relatedmortality on each ramet and the proportion of leafarea lost from ramet mortality were assessed (asdescribed above) for each tree sampled for SNF.Leaf punches from the recently-matured leaveswere taken in the field from each genet for whichSNF was measured, and analyzed for specific leafarea (SLA = cm2 g 1) after drying for 48 hat 40oCin the laboratory. Total N and 8l5N of leaves weredetermined by mass spectrometry as describedabove, and total P was determined colorimetricallyfollowing perchloric acid digestion.

A. tenuifolia ramets greater than 1 m in height weremeasured at 1.37 m (DBH). Surveys were used todevelop plot-level estimates of disease incidence(percentage of live ramets colonized) and disease-related mortality (% ramets dead with canker, %basal area dead with canker, %, canopy deathassociated with canker). Canker incidence on eachramet. based on visible presence of conidiomata orascomata (these forms of fruiting are difficult todistinguish in the field) and discrete canker mar-gins, was quantified as follows. For ramets smallerthan 4 cm diameter, ramets were scored as eitherwith or without canker. For ramets larger than4 cm diameter, main ramets and side brancheswere scored as either with or without canker. Dis-ease-related mortality of ramets was quantified byscoring recently dead ramets in both size classesas either with or without canker. To accuratelycount individual ramets. plots were roped-off into5 x 10m2 sectors. Within each sector, the percentloss of leaf area due to canker-related ramet mor-tality was assessed visually, and values (0-100%canopy loss) were averaged across sectors to pro-vide one plot-level estimate. Ramets without dis-cernible canker colonization but with dead distalbranches were scored separately. All ramets largerthan 4 cm diameter were labeled with a uniquelynumbered aluminum tag affixed with wire forlater determination of the rates of disease-relatedmortality.

N-Fixation Rate

Ten genets were selected within each plot for thedetermination of N-fixation rate. Because wewanted to assess how the incidence of canker dis-ease was affecting Nvfixation rate at the plant(genet) level. we arbitrarily selected individualgenets that spanned the range of disease incidencewithin the genet (proportion of live ramets colo-nized) and disease-related mortality within thegenet (proportion of dead ramets with canker)within the plot. N-lixation rate of each genet wasmeasured using a 15N2 uptake method (Andersonand others 2004). Approximately 2.5 g of freshnodule with attached fine roots were harvestedfrom beneath each genet, placed in a 60 ml poly-ethylene syringe fitted with a septum, and buriedin the litter layer to maintain ambient tempera-tures. Ten milliliter of 99 at. % 15N2 (Isotec Inc..Miamisburg, Ohio, USA) was then added to thesyringe to produce an incubation atmosphere of20.3% 15N2. Immediately after 15N2 addition, a15-ml sample of the incubation atmosphere wasremoved to provide a quantitative measure of atom

Nodule Biomass and N InputsBecause the density of individual A. tenuifoliagenets and the number of ramets per genet variedsignificantly among regions, it was important toselect a method for quantifying nodule biomassthat could be applied across all plots. We firstquantified average nodule biomass beneath aldercanopies (g nodule biomass per m2 canopy), andthen scaled values to the stand-level by developingregion-specific relationships between genet basaland canopy areas. Five individual genets thatspanned a range of genet size, disease incidence,and disease-related mortality were selected forsampling at each plot. The level of disease inci-dence was an important criterion because pre-liminary sampling revealed a large number of deadnodules beneath genets with high incidence ofdisease and associated mortality. We predicted thatestimation of the percent decline in live nodulebiomass would be a useful way of estimatingthe extent to which canker infection negativelyimpacted whole stand N-fixation inputs. The can-opy diameter of each genet was estimated, andthree evenly spaced concentric bands (rings) weremarked on the soil surface beneath the canopy.Multiple soil cores (15.2 cm diameter x 15.2 cmdeep) were sampled from random locations withineach band. More cores were sampled from theouter bands to sample approximately 10% of thesurface area from each band. Soil cores were storedin plastic bags at 4°C for up to 2 days prior tosorting. Nodules clusters were washed free of soiland sorted into live (by 1 cm size classes) and deadbiomass based on color. Unlike roots, live aldernodules are easy to identify due to their charac-teristic bright orange color, whereas dead nodulesare black. Sorted nodules were then dried andweighed. Attenuation curves (decrease in biomassaway from the central ramet cluster) were estab-lished for each genet, and relationships betweenbasal area and nodule biomass were determinedacross all three plots for each region. Total nodulebiomass (g m- 2) was then estimated for each plotfrom values of basal area. N inputs through N2fixation at the plot level were determined bymultiplying average SNF (across ten genets) andnodule biomass for each plot.

Statistical AnalysesVariations among sites in site characteristics, diseaseincidence, canker-related mortality, and parame-ters related to N fixation were analyzed usingANOVA (PROC GLM) (SAS, 2002). Data were testedfor normality and homogeneity of variances, and

square-root or loglo (X + 1) transformed, or rankedand reanalyzed where necessary to meet ANOVAassumptions. Linear and nonlinear regressions wereused to characterize relationships among plantgrowth traits, stand characteristics, and diseaseincidence and mortality. Statistical significance was

also considered. Unless otherwise stated. data nre-

RESULTS

Canker SurveyThe percentage of ramets either colonized or dead

P = 0.07) (Figure I). The incidence of canker dis-ease on live ramets was also significantly higher

strongly influenced by the high disease incidence ofthe large proportion of smaller-diameter ramets. forwhich the typical diffuse canker disease symptomswere more difficult to characterize than for largerramets. If we only consider the percentage of liveramets larger than 4 cm diameter with main ramet

2% at Quartz Creek, Eagle River, and TananaRiver, respectively (F2,6 = 3.1, P = 0.12). Similarly,the proportion of all ramets larger than 4 cmdiameter dead with canker averaged

related mortality of those genets. In other words,variation in SNF among sites or replicate standswithin sites could not be attributed to variation indisease-related mortality (% basal area dead withcanker, or %, canopy dead with canker) or diseaseincidence (% live ramets colonized) (ANOYA datanot shown). Neither did we detect any effect on SNFby wooly alder sawfly damage, which was onlypresent at Eagle River, but where leaf area loss fromsawfly among genets sampled for N fixation aver-aged 18 ± 5%) (range 1-75%).

29 ± 9, and 6 ± 2 % at the three sites, respectively(F2,6 = 2.8, ns). Across sites, the percentage of theselarge ramets (> 4 cm diameter) that were deadwith canker accounted for 55% of the variation intotal dead basal area (P < 0.05) and 94% of vari-ation in the percentage of total basal area dead withcanker (P < 0.0001).

We established a relationship between canopydeath and dead basal area which also shows higherdisease-related mortality for Eagle River and QuartzCreek. At Quartz Creek, Eagle River, and theTanana River, the percentages of basal area deadwith canker averaged 28 ± 11, 30 ± 8, and 11 ± 4%(F2,c, = 1.7, ns). and the percentage of canopy lossassociated with canker averaged 20 ± 8, 15 ± 8,and 2 ± 1% (F2,c, = 1.9, ns) at the three sites,respectively, The interrelationship between thesetwo metrics provides a simple means for predictingdisease-related basal area loss (%DEADBA) from%canopy loss ('%CANLOSS), a parameter thatcan be rapidly assessed in the field (%DEADBA =1.07 x %CANLOSS + 3,03; r2 = 0,84, P < 0,0001),A complete characterization of disease incidence anddisease-related mortality by diameter class for thethree sites shows that the proportions of size classeslarger than 4 cm diameter that were dead with can-ker tended to be higher for Eagle River and QuartzCreek relative to Tanana River plots (Figure 1),

N FixationN-fixation rates at Quartz Creek (5,52 ± 1.23 umolN g -1 h -1 were significantly less than those alongthe Eagle River (8,57± 0,78 umol N g-1 h-1) andTanana River (9,75 ± 1.05 umol N g-1 h-1), whichdid not differ (F2,7,8, = 5,05, P < 0.05). SNF waspositively correlated with soil temperature acrosssites (r2 = 0.12, P < 0.001, n = 90), but unrelatedto other soil chemical characteristics or plant traits(data not shown). Data on soil chemical character-istics and A. tenuifolia leaf traits from the three studyregions are shown in Supplementary Material:Soil and Leaf Chemistry, and can be found on theBonanza Creek LTER website (http://www.lter.nal.edu/).

Incidence of canker disease and related mortalityvaried substantially among the 90 genets measuredfor N-fixation rate across sites, and the relation-ship between % basal area dead and % canopy lossgenerated for these genets (Figure 3;) was similar tothat from plot-level measurements mentionedabove. Despite this variation in disease incidenceand related mortality, we found no evidence thatN-fixation rate per gram nodule of individual gen-ets was influenced by the incidence of disease or

Nodule BiomassLive nodule biomass beneath A, tenuifolia canopiesaveraged 28.0 ± 5.9, 22.0 ± 4.4, and 36.4 ± 2.0g m-2 at Quartz Creek, Eagle River, and TananaRiver sites, respectively (Figure 4A), The propor-tion of nodule biomass dead at Tanana River(37.7 ± 3,5%) and Eagle River (42.9 ± 6.4%) wassignificantly greater than that at Quartz Creek(25.6 ± 3.4%) (F2,33 = 3.66, P < 0.05). TananaRiver sites had significantly higher ratios of canopynodule biomass to basal area relative to either ofthe other sites, for ratios expressed as either total(F2,33 = 3.67, P < 0.05) or live components (F2,33 =3.14, P < 0.05). There were also notable differ-ences in the size distributions of live nodulesamong regions, with Tanana River sites havingsignificantly higher proportion of nodules in smal-ler size classes relative to the other two regions(Figure B).

By design, individual genets chosen for nodulebiomass sampling had a wide range of incidence ofcanker disease and related mortality both withinand among sites. For example, the coefficient ofvariation in % basal area dead, % canopy loss, and% incidence of main ramet canker all averagedover 100% among the 12 genets sampled withineach region. Across all sites, live and dead nodulebiomass were correlated with a number of mea-sures of canker disease incidence and related mor-tality, suggesting that higher incidences of thedisease led to declines in live nodule biomass at allsites. For example, genets showing increasing per-centages of basal area dead with canker also tendedto have more dead nodule biomass as a fraction oftotal nodule biomass (r2 = 0.10, P = 0.09). Simi-larly, increased canopy loss associated with canker(%) tended to be inversely correlated with livenodule biomass (g m-2) across regions (r = - 0,30,P = 0.08). The most useful relationship for esti-mating the negative effects of canker diseaseon nodule biomass was the inverse correlationbetween live nodule biomass (LlVENOD, g m-2)

N-Fixation InputsWe developed negative exponential relationshipsbetween canopy area (m2

) and basal area (cm2) or

individual shrubs for each region, which whenmultiplied by total stand basal area produced

309 m2 alder canopy cover per hectare for QuartzCreek, Eagle River, and Tanana River stands,respectively. This allowed stand-level estimates oflive nodule biomass, which averaged

per hectare for the three stands, respectively. Wethen estimated daily rates of N input for each of thenine stands as the product of N-lixation rates andstand-level nodule biomass, assuming fluxes wereconstant over a 24- h period, as we have observed(R.W. Ruess, unpublished data) and has beenshown elsewhere at high latitudes (Weisz andSinclair 1988; Huss-Danelland others 1992). Daily

Quartz Creek, Eagle River, and Tanana river,respectively, were extrapolated to the growingseason using a simple step function of plant growthand N fixation based on our previous studies (Uli-assi and Ruess; 2002; Anderson and others 2004).This function assumes that N-fixation rates arehalf-maximal from 20-31 May, maximal from June1 to August 15, half-maximal for the last 2 weeksof August, and quarter-maximal for the first2 weeks of September. This generated N-fixation

N ha -1 y -1 for Quartz Creek, Eagle River, andTanana River sites, respectively.

To determine the extent to which canker diseaseincidence and related mortality has impactedN-fixation inputs, we predicted what live nodulebiomass would have been in the absence of cankerdisease from region-specific negative exponentialrelationships between LlVENOD and %CANK (seeabove). Estimates suggest that live nodule biomassper m2 canopy has been reduced 24.4, 25.3, and22.5% by the canker disease and associatedmortality at Quartz Creek, Eagle River, and TananaRiver stands, respectively. Such declines in nodulebiomass would translate to N input reductions of 8,16, and 33 kg N ha -1 y-1 for the three regions,respectively. Disease-free stands at Quartz Creek,Eagle River, and Tanana River would be predicted

therefore to be accumulating 30, 58, and 140 kgN ha -1 y-1 through N fixation ,respectively. Weview the translation of these values to declines inN-fixation inputs to these stands as conservative,because the percentages of dead nodule biomass thatwe found far exceed values we remember seeing adecade ago when we conducted similar N inputestimations along the Tanana River.

DISCUSSION

Spread of Canker on Thin-Leaf Alderin Interior AlaskaOur data show tha tthe spread of fungal stem cankerdisease in A. tenuifolia is both recent and widespreadthroughout interior and south-central Alaska, andhas led to significant ramet mortality and associateddeclines in ecosystem N inputs derived from N fix-ation. Although there are no previous records ofcanker epidemics on A. tenuifolia in Alaska, V. mel-anodiscus is most likely native to interior Alaska,because V. alni (a synonym for V. melanodiscus) islisted from old Alaskan fungal collections (Cash(1953; Conners 1967 "). One of the most intriguingquestions is whether the high disease incidence andwidespread canker-related ramet mortality that wedocumented is new to interior Alaska, or is simply areturn of a natural epidemic that has occurredeither regularly or episodically for centuries. Ifsimilar epidemics have occurred previously, thenlegacies or periodic interruptions to N-fixation in-puts during early-successional development maypotentially be found in growth or chemical signa-tures of annual rings from cohorts of long-livedspecies (such as white spruce) recruited at that time.However, if the widespread disease and mortality ofA. tenuifolia is a recent phenomenon, then associ-ated declines in N-fixation inputs may have novelimplications for successional and landscape devel-opment that will ultimately depend on the rate ofalder reestablishment.

The unusually hot and dry summer climatethroughout Alaska during the current decade mayhave contributed to the recent increase in incidenceand severity of canker disease on A. tenuifolia.Interactions between drought and fungal diseasesare a growing concern, particularly throughoutEurope and North America where climate warminghas led to unusually warm and dry conditions overthe past decade (Desprez-Loustau and others

2006; Bigler and others 2007;Worrall and others2008). Predisposition by water stress may involve

both improved growth conditions for the fungi inwater-stressed plants or a decrease in active disease

resistance when plants are water stressed (Guyonand others 1':l'll); Desprez-Loustau and others}O(lh). Cytospora canker fungi are characteristicallyopportunistic pathogens that attack stressed ordamaged hosts, and water stress is thought .to playan integral role in the initiation and development ofCytospora canker diseases of hardwoods. Droughtstress and wounding are often combined withexperimentally initiate infection (Bloomberg andFarris I. fi(,); Filip and others ! ')():); Kepley andJacobi ~~(H)O), and critical levels of bark moisturecontent have been correlated with increased sus-ceptibility to Cytospora canker fungi (Bloomberg19(2). Worrall and others (20!),'\) reported _thatdrought was likely an important driver of Cytosporacanker contributions to mortality of tremblingaspen throughout the Rocky Mountain States.Drought has similarly been implicated in facilitatingCytospora canker-induced dieback of A. tenuifoliathroughout Oregon, Washington, and Colorado(Filip and others i . Worrall and Adams .!!ii

Branch damage, severe defoliation, and interac-tions with other pathogens may also contribute todisease initiation or disease severity. ThroughoutAlaska, A. tenuifolia often suffers extensive branchdamage during winter when heavy snow can flat-ten even the largest shrubs, particularly after rametshave accumulated heavy layers of frost (personalobservation). Wounds are classically known tocreate infection courts for canker pathogens, andnatural modes of stem wounding or branch damagehave been implicated in canker-related rametmortality in Alnus incana and Alnus g/util1osa (Mor-icca 2002). It also appears that stress induced bydefoliator outbreaks from native and non-nativesawflies may have contributed to the canker spreadalong the Eagle River, the only region showing anysawfly damage in this study. For example, amongthe shrubs sampled for nodule biomass along theEagle River, the proportion of canopy defoliated bysawfly was positively correlated with the proportionof the canopy dead from canker r2 = 0.74, P <0.001). Such dramatic changes in host conditionmay allow disease to develop rapidly. One possi-bility is that a shift in host physiology can result indisease development by releasing a quiescentpathogen from latency, a relatively long period ofhost colonization without damage to the host(Stanosz and others 2000). Several Cytospora spe-cies have previously been isolated from asymp-tomatic hosts (Chapela 1989; Adams and others2005), suggesting the potential for vigorous hoststo impose latency. Disease progression can alsobe facilitated by weakly parasitic fungi, whichhave characteristics that allow quick invasion of

Effects of Canker on N-Fixation InputsWe estimate thai during the 2005 growing season,canker disease incidence and canker-related mor-tality reduced N-fixation inputs by 8, 16, and 33 kgN ha -1 y -1 within alder stands along Quartz Creek,Eagle River, and the Tanana River, respectively.Live nodule biomass scaled inversely with thepercentage of ramets dead or live with main rametcanker, resulting in disease-related reductions inlive nodule biomass per m 2 canopy of 24.4, 25.3,and 22.5% at the three study regions, respectively.N-fixing plants typically up- or down-regulatenodule production in response to factors affectingplant growth and plant N demand. Growth andallocation to nodules in response to ratios ofavailable soil N to P are well described (Wall andothers 2000, 2003; Uliassi and Ruess 2002; Binkleyand others 2003; Valverde and Wall 2003; Jiaand others 2004; Gokkaya and others 2006), butfewer studies have examined nodule mortality inresponse to stress in natural stands. Increasedmortality and turnover of nodule biomass has beenshown for neotropical leguminous trees that aremanaged with pruning (Nygren and Ramirez

1995), similar to the response of fine productionand lifespan to aboveground browsing (Ruess andothers 1998). There is also some evidence thatnodule biomass decreases and nodule lifespanincreases with stand age (Sharma and Anibasht

nodule clustersnodule biomass (32.8%) in this study was notablygreater than for similarly sized alders sampled alongthe Tanana River in 1994 (24.0%) prior to thecanker outbreak (Uliassi and Ruess 2002). Thatprevious study found that 20.3% of nodule biomasswas in clusters 3 cm diameter or larger, whereas wefound only 1 % of nodule biomass in these oldersize classes, suggesting that nodule lifespan hasdeclined appreciably since the increased incidenceof disease and associated widespread mortality.

One explanation for the high disease incidencewithin early-successional stands along the TananaRiver and Eagle River is simply the high density ofalder ramets in these stands. A positive correla-tion between host density and disease incidence istypical of many plant fungal diseases (Gilbert

2002), and often results directly from increasingpathogen-host encounter probabilities. Additionalindirect effects include density-dependent physio-logical plant responses that increase host vulner-ability to infection, and the effects of increasedhost density on improving environmental condi-tions favorable to the pathogen (Burdon andChilvers 1982; Burdon and others 1989; Gilbert

2002). Disease incidence and related mortality waspositively correlated with stand density across allstands we surveyed in 2005 (Figure 5), suggestingthat canker epidemics may be facilitated by peri-odic alder recruitment events.

weakened host tissue and can facilitate diseaseprogression (Kehr 1992). Although the ability of V.melanodiscus to incite cankers on A. tenuifolia isconfirmed (Stanosz and others 2008), other closely-related fungi have been found in association withdiffuse cankers and dieback on alders in Alaska(USDA 2008). The role of biotic and abiotic factorsin the initiation and development of the diffusecanker disease and dieback process is currentlybeing investigated.

Although we found lower canker-related mor-tality along the Tanana River, the higher incidenceof disease in younger-aged ramets suggests thatdemographics of the local population may play arole in disease development. Tanana River standsalso had a much higher proportion of root nodulesin smaller size classes, suggesting there is a morerapid turnover of nodules along the Tanana River.Higher nodule turnover rates could also simply be afunction of the higher relative growth rate of A.tenuifolia along the Tanana River relative to theother two regions, as suggested by higher leaf N,thinner leaves, higher N-fixation rates, and higherapparent allocation to nodules in the Tanana Riverplots. However, the average contribution of young

diameter) to total live

1986; Pearson and Vitousek 2001). Although wepreviously found no significant effects of defolia-tion on allocation to nodule biomass in green-house-grown A.tenuifolia seedlings (Ruess andothers 2006, increased levels of sawfly defoliationamong plots along the Eagle River were associatedwith increased levels of canker-related rametmortality in this study. Given the modular growthof multiple ramets arising at the ground surfacefrom a singlegenet of A. tenuifolia growing in thefield, it is logical that root and nodule diebackwould scale to the proportion of genet basalarea death related to canker, and differ from theresponse of single-ramet seedlings grown in thegreenhouse.

Our estimates of potential N-fixation inputs toalder stands unaffected by canker disease along theTanana River (140 kg N ha -1 y-1: are similar toprevious st udies conducted within the same region.Using a mass balance approach accounting for allvegetation and soil to a depth of 1 m. Van Cleveand others (1971) reported an average total systemN accretion of 156 kg N ha-1 y-1 (range 75-362 kgN ha-1 y -1) [or alder thickets at 5, 15, and 20 yearsfrom initial surface establishment. Uliassi andRuess (2002) estimated N-fixation inputs of 59 kgN ha -1 y -1 for A. tenuifolia stands with stem den-

sities (mean 6,667 stems ha-1) and basal areas

(mean 20.5 m2 ha -1) that were less than half thoseof this study. Our estimate of a 24% decline inN-fixation inputs during 2005 by A. tenuifolia alongthe Tanana River is a maximal value, because itassumes that all dead nodule biomass resulted fromcanker-induced ramet mortality. However, ourexperience is that dead nodules are relatively rarein early-successional alder stands.

Consequences for Ecosystem Functionand Landscape ChangeThe influence of N-fixing shrub encroachment onbiodiversity ancl ecosystem function is a topic ofconsiderable interest and concern (Vitousek andWalker 1989; Hibbard and others 2001 : Caldwell.: 2004; Rice and others 2004; Baer and others 2006;Hughes and others 2006). Yet the consequences forloss of native N fixers on ecosystem function arepoorly documented. Mass accumulation estimatessuggest that up to 70% of ecosystem N accumu-lated over the Tanana floodplain 200-year succes-sional sequence is fixed by A. tenuifolia within thefirst 30 years of succession (Van Cleve and others

1971 , 1991). Paleoecological records indicate thatwatershed biogeochemistry and aquatic trophicdynamics were substantially influenced by alder

expansion throughout southern and northwesternAlaska during the early- to mid-Holocene (Oswaldand others 1999); Hu and others 2001).

During the 2008 growing season, we resurveyedthe Tanana River plots used for this study andcharacterized the incidence and related mortalityfrom canker disease on all ramets. including thoselarger than 4 cm diameter, all of which were indi-vidually tagged in 2005. Of the ramets larger than4 cm diameter having main ramet canker in 2005,74% are now dead; and for those without mainramet canker in 2005, 25% have developed mainramet canker, and 8% are dead. Moreover,recruitment of new individuals is nonexistentwithin alder thickets (data not shown). Thus, it ishighly likely that N-fixation inputs have declinedsignificantly below what we estimated for 2005.Increase in disease incidence and disease-relatedmortality have also continued, but at a slower pace,in mid-successional balsam poplar stands (data notshown), where A. tenuifolia forms an understoryless-dense than in the younger alder stage (Viereckand others 1993). However, vegetative propagationfrom downed ramets (principally snowfall induced)is extensive in these stands, ostensibly due tohigher light availability. Additional data suggestthat much of the young, sparsely-distributed A.tenuifolia recruiting on newly-formed sandbars(< 10 years of age) also has lower levels of diseaseincidence and related mortality than that wasfound for dense alder thickets described in thisstudy (Nossov 2008 ). Thus, although the densealder thickets along the Tanana River appear to bedying out, it remains to be seen whether A.tenui-folia in less-dense stands can persist through thecurrent canker disease epidemic.

How the loss of dense alder stands will influenceplant growth and successional development alongriver terraces throughout interior Alaska is ofcourse unknown. A complete die-out of thin-leafalder would clearly have serious consequences forlong-term ecosystem productivity. However, asmentioned above, we are uncertain whether thecurrent epidemic represents a threshold shift in Ncycling dynamics (Chapin and others 2006), ormerely a return of an infrequent co-occurrence ofevents. Because alder has strong effects on Pmobilization and cycling (Giardina and others1995; Binkley and others 2000; Mitchell and Ruesssubmitted), even a temporary loss of alder willlikely have effects on biogeochemical cyclingbeyond the immediate influence on soil N stocks.Analogs for how the loss of alder might influenceecosystem structure and function can be foundalong interior Alaskan river corridors where low

moose population densities increase the ratio ofwillows to alder (Butler and others 2007. We arecurrently studying the influence of alder on bio-geochemistry and stand development at a regionalscale of varying moose densities to understand theimplications for permanent disturbance-mediatedshifts in alder abundance.

ACKNOWLEDGEMENTS

We wish to thank a number of enthusiastic, hard-working undergraduate students who participatedwith field studies and laboratory analyses critical tothe success of this project. These include AnnaMarx, Nils Petersen, Keane Richards, Eric Robinson,Amanda Roberson, and Dorothy Walker. We areindebted to L. Oliver at the UAF Forest Soils Labo-ratory for providing her expertise and resources formass spectrometry analyses. Two anonymousreviewers improved the manuscript substantially.Funding for the research was provided by the USForest Service, Region 10, State and Private For-estry, Anchorage, Alaska, by the Bonanza CreekLong- Term Ecological Research program (fundedjointly by NSF Grant DEB-0620579 and USDAForest Service, Pacific Northwest Research StationGrant PNWOI-JVI1261952-231), and by NSF GrantDEB-0641033 to R.W. Ruess.

REFERENCES

Adams GC, Wingtield MJ, Common R, Roux J. 2005. Phyloge-netic relationships and morphology of Cytospora species andrelated teleomorphs (Ascomycota, Diaporthales. Yalsaceae)from Eucalyptus-preface. Stud Mycol 52:1-144.

Allen CD. 2007. Interactions across spatial scales among forestdieback. fire, and erosion in northern New Mexico landscapes.Ecosystems 10:797-808.

Anderson MD, Ruess RW, UIiassi DD, Mitchell JS. 2004. Esti-mating No fixation in two species of Alnus in interior Alaskausing acetylene reduction and 15N2 uptake. Ecoscience11:102-12.

Baer SG, Church J.M., Williard KWJ, Groninger JW. 2006.Changes in intra system N cycling from N2fixing shrubencroachment in grassland: multiple positive feedbacks. AgricEcosyst Environ 115:174-82.

Berg EE, Henry JD, Fastie CL, DeVolder AD, Matsuoka SM.2006. Spruce beetle outbreaks on the Kenai Peninsula, Alaska,and Kluane National Park and Reserve, Yukon Territory:relationship to summer temperatures and regional differencesin disturbance regimes. For Ecol Manage 227:2 1 9-32.

Bigler C, Gavin DG, Gunning C, Veblen TT. 2007. Droughtinduces lagged tree mortality in a subalpine forest in theRocky Mountains. Oikos 116:1893-994.

Binkley D, Giardina C, Bashkin MA. 2000. Soil phosphoruspools and supply under the influence of Eucalyptus salignaand nitrogen-fixing Albizia falcataria. For Ecol Manage128:241-7.

Binkley D, Senock R, Cromack K Jr. 2003. Phosphorus limitationon nitrogen fixation by Facaltaria seedlings. For Ecol Manage186:171-6.

Bloomberg W. 1962. Cytospora canker of poplars: the moisturerelations and anatomy of the host. Can J Bot 40: 1281-9 3.

Bloomberg W, Farris S. 1963. Cytospora canker of poplars: barkwounding in relation to canker development. Can J Bot41:303-10.

Burdon Ll. Chilvers GA. 1982. 110st density as a factor ill plantdisease ecology. Annu Rev Phytopathol 20:143-(,6.

Burdon JJ, Jarosz AM, Kirby Gc. 19R9. Pattern and patchinessin plant-pathogen interactions-causes and consequences.Annu Rev Ecol Syst 20: 119-36.

Butler LG, Kielland K. 2008. Acceleration of vegetation turnoverand element cycling by mammalian herbivory in riparianecosystems . J Ecol 96: I 36-44.

Butler LG, Kielland K, Scott Rupp T, Hanley TA. 2007. Interac-tive controls of herbivory and fluvial dynamics Oil landscapevegetation patterns on the Tanana River Hoodplaiu. interiorAlaska. J Biogeogr 14: 1622-3 1.

Caldwell BA. 2004. Effects of invasive scotch broom on soilproperties in a Pacific coastal prairie soil. Appl Soil Ecol32:149-52.

Cash E. 1953. A checklist 01 Alaskan fungi. Plant Dis Rep Suppl219:1-70.

Castello JD. Leopold OJ, Smallidge P.J. 1995. Pathogens, pat-terns, and processes in forest ecosystems. Pathogens influenceand are influenced by forest development and landscapecharacteristics. Bioscience 45: 16-24.

Chakraborty S, Datta S. 20m. How will plant pathogens adapt tohost plant resistance at elevated CO2 under a changing cli-mate? New Phytol 159:730--42.

Chapela IH. 1989. Fungi in healthy stems and branches ofAmerican beech and aspen-a comparative study. New Phytol113:65-75.

Chapin FS III, Oswood MW, Van Cleve K, Viereck LA, verbylaDL, Eds. 2006. Alaska's changing boreal forest. New York:Oxford University Press.

Cobb R, Orwig D, Currie S. 2006. Decomposition of green foliagein eastern hemlock forests of southern New England impactedby hemlock woolly adelgid infestations. Can J For Res36:1331-41.

Conners IL 1967. An annotated index of plant diseases in Canadaand fungi recorded on plants in Alaska, Canada, and Green-land. Ottawa, Canada: Canada Dc-partrne-ru of Agriculture,Research Branch Canada Department of Agriculture, p 1- 381

Desprcz-Loustau ML, Marcais B, Nagelcisen L-M, Pion D, Van-nini A. 2006. lnteractive effects of drought and pathogens inforest trees. Ann For Sci 65:597-612.

Ellison AM, Bank MS, Clinton BD, Colburn EA, Ellint K, FordCR, Foster DR, Kloeppel BD), Knoepp JD Lovett GM, MohanJ, Orwig DA, Rodcnhuuse NL Sobczak WV, Stinson KA,Stone JK, Swan CM, Thompson J, Von Holle B, Webster .m.2005. Loss of foundation species: consequences for thestructure and dynamics of forested ecosystems. Front EcolEnviron 3:479-86.

Erickson JE, Stanosz GR, Kruger EL. 2004. Photosyntheticconsequences of Marssonina leaf spot differ between twopoplar hybrids. New Phytol 161:577-83.

Faeth SH Hammon KE. 1997. Fungal endophytes in oak trees:experimental analyst's of interactions with leafminers. Ecol-ogy 78:820-7.

Filip GM. Parks CA. St arr GL. 1992. Incidence of wound-asso-cialed infection by Cytospora sp. in mountain alder, red-osierdogwood, and black hawthorn in Oregon. Northwest Sci66: 194-8.

Giardina CP. Huffman S, Binkley D, Caldwell BA. 1995. Aldersincrease soil phosphorus availability in J Douglas-fir planta-tion. Can J For Res 25:1652-7.

Gilbert GS. 2002. Evolutionary ecology of plant diseases innatural ecosystems. Annu Rev Phyropathol 40: 13-43.

Gokkaya K. Hurd TM. Ravnal DJ. 2006. Symbiont nitrogenase.alder growth. and soil nitrate response to phosphorus addi-lion in alder (Alnus incana ssp. rugosa) wetlands of theAdirondack Mountains. New York State, USA. Environ ExpHot 55:97-109.

Guyon JC. Jacobi WR. Mclntyre GA. 1996. Effects of environ-mental stress on the development of Cytospora canker of aspen.Plant Dis 80: I 320-6. '

Hansell EM. Goheen EM. 2000. Phellinus weirii and other na-tive root pathogens as determinants of forest structure andprocess in western North America. Annu Rev Phytopathol38:515-39.

Hibbard KA. Archer S. Schimel DS. Valentine DW. 2001. Bio-geochemical changes accompanving woody plant encroach-ment in a subtropical Savanna. Ecology 82:1999-2011.

Hu F-S. Finney BP, Brubaker LB.2001. Effects of Holocene Alnusexpansion on aquatic productivty. nitrogen cycling and soildevelopment in Southwestern Alaska. Ecosystems 4:358-68.

Hughes RF. Archer SR. Asncr GP. Wessman CA, McMurtry C.Nelson JIM. Ansley RJ. 2006. Changes in aboveground pri-mary production and carbon and nitrogen pools accompany-ing woody plant encroachment in a temperate savanna. GlobChang Bioi 12: 1733-47.

Huss-Danell K.Lundquist PO. Ohlsson II. 1992. N2 fixation in ayoung Alnus incana stand. based on seasonal and diurnalvariation in whole plant nurogenase activity. Can J Bot70: 1517--44.

Jenkins .I. Aher .I, Canham C. 1999. Hemlock woolly adelgidimpacls on communitv structure and N cycling rates in east-ern hemlock forests. Can J For Res 29:630-45.

Jenkins MA. Jose S. White PS. 2007. Impacts of an exotic diseaseand vegetation change 0n foliar calcium cycling in Appala-chian forests. Ecol Appl 17:869-81.

Jia Y. Gray VM, Straker CJ. 2004. The influcncc of Rhizobiumand arhuscular mycorrhizal fungi on nitrogen and phosphorusaccumulation by vicia faba. Ann BOt 94:251-8.

Kehr R. 1992. Pezicula canker of Querus rubra L., caused byPezicula cinnamomea (DC.) Sacc. Part II. Morphology andbiology of the causal agent. Eur J For Pathol 22:29-40.

Kepley J. Jacobi WH. 2000. Pathogenicity of Cytospora fungi onsix hardwood species J Arboric 26: 326-33.

Lajtha K. Driscoll CT, Janell WM, Elliott ET. 1999. Soil phos-phorus: characterization and total element analysis, In: Rob-ertson GF. Coleman DC, Bledsoe CS, Sullins P. Ecls. Standardsoil methods for long-term ecological research. New York:Oxford University Press. p 115-42,

Ligluloo: DC. Whitford WG. 1990. Phytophagous insects en-hance nitrogen flux in a desert rreosotebush community.Occologia 1l2: IS-25.

Malmstrom CM. Raffa KF. 2000. Biotic disturbance agents in thehorcal forest: considerations for vegetation change models.Glob Chang Kiol 6: 35--48.

Mitchell JS, Ruess RW. Submitted. Patterns of nitrogen inputs byAlnus viridis spp, fruticosa to a secondary successional chron-osequence in interior Alaska. Biogeochemistry.

Moricca S. 2002. Phoniopsis alnea, the cause of dieback of blackalder in Italy. Plant Pathnl 51 :755-64.

Mulder CPH. Roy EA. Gusewell S. 200R. Herbivores andpathogens on Alnus viridis subsp. fruticosa in interior Alaska:effects of leaf, tree and neighbour characteristics on damagelevels. Oecologia Botany 86:408-21.

Nossov D, 2008. Community, nopulation. and growth dynamicsof thin1eaf alder (Alnlls tenuifolia: implications for nutrientcycling on an interior Alaskan floodplain. Fairbanks, AK:University of Alaska.

Nygren P, Ramirez C. 1995. Production and turnover of N2 fixingnodules in relation to foliage development in periodicallypruned Erythrina poeppiaiana (Leguminosac) trees, For EcolManage 73:59-73.

Oswald W, Brubaker L. Anderson P. 1999. Late Quaternaryvegetational history of the Howard Pass Area, northwesternAlaska. Can J Bot 77:570-81.

Parker TJ. Clancy KM, Mathiasen RL. 2006. Interactions amongfire, insects and pathogens in coniferous forests of the interiorwestern United States and Canada. Agric For Entomol 8:167-89.

Pearson HL, Vitousek PM. 2001. Stand dynamics, nitrogenaccumulation. and symbiotic nitrogen fixation in regeneratingstands of Acacia koa. Ecol Appl II: 1381-94.

Reynolds BC, Hunter MD, 2001. Responses of soil respiration.soil nutrients, and lurer decomposition to inputs from canopyherbivores. Soil Bioi Biochem 33:1641-52.

Rice SK. Westerman B. Federici R. 2004. Impacts of theexotic, nitrogen-fixing black locust (Robinia pscudoacacias onnitrogen-cycling in a pine-oak ecosystem. Plant Ecol 174:97-104.

Ruess RW, Anderson MD. Mitchell JS, Mcf'arland JW. 2006.Effects of defoliation on growth and N-fixation in Alnus len-uiiolia: consequences for changing disturbance regimes at highlatitudes. Pcoscteuce l3:402-12.

Ruess RW, Hendrick RL. Bryant JP. 1998. Regulation of fine rootdynamics by mammalian browsers in early successionalAlaskan taiga forests. Ecology 79:2706-20.

Scherm H. 2004. Climate change: can we predict the impacts onplant pathology and pest management? Can J Plant Pathol26:267-73.

Sharma E, Ambasht RS. 1986. Root nodule age-class t ransition.production and decomposition in an age sequence of Alnusnepalensis plantation stands in the Eastern Himalayas . .J ApplEcol 23:689-701.

Snyder C, Young J, Leman D, Smith D. 2002. Influence ofeastern hemlock (Tsliga canadensis) forests on aquatic inver-tebrate assemblages in headwater streams. Can J Fish AquatSci 59:262-75,

Stadler B, Muller T. Orwig D. 2006. The ecology of energy andnutrient fluxes in hemlock forests invaded by hemlock woollyadelgid. Ecology 87: 1792-805.

Stanosz GR, Blodgett J. Smith D. Kruger EL. 2000. Water stressand Sphtteropsis sapinea as a latent pathogen of red pine seed-lings. New Phytol 149: 531-8.

Stanosz GR. Trummer LM. Rohrs-Richey JK<,Adams GC, Worrall.l.l. 2008. Response of Alnus tennifolia In inoculation with vatsamelanodiscus. Phytopathology 98:S I 50.

Tainter FH.O'Brien JG, Hernandez A, Orozco F, Rebolledo O.2000. Phytophthora cinnamomi as a cause of oak mortality inthe state of Colima, Mexico. Plant Dis 84:394-8.

Tingley MW, Orwig DA, Field R, Motzkin G. 2002. Avianresponse to removal of a forest dominant: consequences ofhemlock woolly adelgid infestations. J Biogeogr 29( 10-11):1505-16.

Uliassi DD. Ruess RW. 2002. Limitations to symbiotic nitrogenfixation in primary succession on the Tanana-River floodplain,Alaska. Ecology 83:88-]03.

USDA. 2008. Forest health conditions in Alaska-2007. Wash-ingron. DC: USDA Ed. Protection Report RIO-PRI8.

Valverde C. Wall LG. 2003. The regulation of nodulation,nitrogen Iixation and ammonium assimilation under a car-bohydrate shortage stress in the Discaria trinervis-Frankiasymbiosis. Plant Soil 254: 155-65.

Van Cleve K, Chapin FS Ill, Dyrness CT, Viereck LA. 1991.Element cycling in taiga forest: state-factor control. Bioscience41:78-88.

Van Cleve K. Viereck LA. Schlentner RL. 1971. Accumulation ofnitrogen in alder (Alnlls) ecosystems near Fairbanks, Alaska.ArCI Alp Res 3:101-14.

Viereck LA. Dyrness CT, Foote MJ. 1993. An overview ofthe vegetation and soils of the floodplain ecosystems of theTanana River, interior Alaska. Can J For Res 23:889-98.

Vuousek PM. Walker LR. 1989. Biological invasion by MyricilFaya in Hawai'i: plant demography, nitrogen fixation, eco-system eilerts. Ec0l Monogr 59:247-65.

Wall LG, Hellsten A, Huss-Danell K. 2000. Nitrogen, phospho-rus, and the ratio between them affect nodulation in Alnusincana and Trifolium pratens. Symbiosis 29:91-105.

Wall LG, Valverde C. Huss-Danell K. 2003.Regulation of nod-ulation in the absence of N2 is different in actinorhizal plantswith different infection pathways. J Exp BOt 54: 1253-8.

Waring RH. Crornack K. Matson PA. Boone RD. Stafford SC.1987. Responses to pathogen-induced disturbance: decem-position, nutrient availability. and tree vigor. Forestry 60:219-27.

Weisz PRo Sinclair TH. 1988. Soybean nodule gas permeabilit v.nitrogen fixation and diurnal cycles in soil temperature. PlantSoil 109:227-34.

Worrall JJ. 2009. Dieback and mortality of Alnus in the southernRocky Mountains, USA. Plant Dis 93:293-8.

Worrall JJ, Adams Gc. 2006. Cytospora canker of Alnus in thesouthern Rocky Mountains, MSA Meeting Abstract. Inoculum(Suppl Mycol) 57:40.

Worrall JJ, Egeland L Eager T, Mask RA . .Iohuson EW, KempPA, Shepperd WD. 2008. Rapid morralirv of Populus tremulo-ides in southwestern Colorado, USA. For Ecol Manage 155:686-96.