Response of the nitrogen-fixing lichen Lobaria pulmonaria to phosphorus , molybdenum , and vanadium

Nitrogen-fixing lichens (cyanolichens) are an important source of nitrogen (N) in Pacific Northwest forests, but limitation of lichen growth by elements essential for N fixation is poorly understood. To investigate how nutrient limitation may affect cyanolichen growth rates, we fertilized a tripartite cyanobacterial lichen (Lobaria pulmonaria) and a green algal non-nitrogen fixing lichen (Usnea longissima) with themicronutrients molybdenum (Mo) and vanadium (V), both known cofactors for enzymes involved in N fixation, and the macronutrient phosphorus (P). We then grew treated lichens in the field for one year in western Oregon, USA. Lichen growth was very rapid for both species and did not differ across treatments, despite a previous demonstration of P-limitation in L. pulmonaria at a nearby location. To reconcile these disparate findings, we analyzed P, Mo, and V concentrations, natural abundance dN isotopes, %N and change in thallus N in Lobaria pulmonaria from both growth experiments. Nitrogen levels in deposition and in lichens could not explain the large difference in growth or P limitation observed between the two studies. Instead, we provide evidence that local differences in P availability may have caused site-specific responses of Lobaria to P fertilization. In the previous experiment, Lobaria had low background levels of P, and treatment with P more than doubled growth. In contrast, Lobaria from the current experiment had much higher background P concentrations, similar to P-treated lichens in the previous experiment, consistent with the idea that ambient variation in P availability influences the degree of P limitation in cyanolichens. We conclude that insufficient P, Mo, and V did not limit the growth of either cyanolichens or chlorolichens at the site of the current experiment. Our findings point to the need to understand landscape-scale variation in P availability to cyanolichens, and its effect on spatial patterns of cyanolichen nutrient limitation and N fixation.


INTRODUCTION
Biological fixation is the primary mechanism by which new nitrogen (N) becomes available to most ecosystems (Vitousek et al. 2002(Vitousek et al. , 2013)).
Nitrogen availability influences primary productivity and carbon exchange, plant growth, community composition, and carbon sequestration in v www.esajournals.orgboth soil and aboveground biomass (Galloway et al. 2003, van Groenigen et al. 2006, LeBauer and Treseder 2008).N-fixers (bacteria, archaea, and cyanobacteria), whether symbiotic or not, contribute available N using the nitrogenase enzyme to convert atmospheric N 2 to NH 3 .N-fixation is often considered to be metabolically expensive, requiring 15 molecules of ATP per molecule of fixed N 2 (Simpson and Burris 1984), and thus large amounts of phosphorus (P).The nitrogenase enzyme also requires iron (Fe), molybdenum (Mo), or vanadium (V) as an essential cofactor (Eady 1996, Einsle et al. 2002, Bellenger et al. 2011).Although a single N-fixer may produce nitrogenase isoenzymes with all three cofactors, Mo-nitrogenase is thought to be used preferentially when Mo is available (Einsle et al. 2002, Bellenger et al. 2011).Recent evidence, however, suggests a strong affinity for V (Darnajoux et al. 2014, Hodkinson et al. 2014).Variation in the availability of Mo, V, P, or other elements may therefore limit the growth and reproduction of N-fixing organisms, as well as N-fixation inputs that shape a range of ecosystem and global processes.
Nitrogen fixing cyanolichens can be bipartite symbioses between a cyanobacteria and a fungus or they can be tripartite symbioses (''cephalolichens'') composed of a fungus and green alga with pockets containing cyanobacteria (cephalodia) throughout the thallus.Cyanolichens are an important source of fixed N, particularly in old forests in wet temperate climates (Pike et al. 1972, Denison 1979, Antoine 2004, Holub and Lajtha 2004), in arctic ecosystems (Gunther 1989), and on freshly exposed substrates where they act as primary colonizers (Crittenden and Kershaw 1978).Recent estimates of N-fixation in cryptogamic ground covers (including lichen, bryophytes, and free-living cyanobacteria and algae) suggest that 46% of biological N fixation on land (49 Tg N uptake per year globally) occurs in these cryptogamic communities (Elbert et al. 2012).Cyanolichen N input can exceed atmospheric deposition and asymbiotic fixation, making it an important long-term N source (Pike et al. 1972, Antoine 2004, Holub and Lajtha 2004, Giesen et al. 2008).Limitation by Mo and P as well as Mo-P co-limitation of asymbiotic N-fixation has been demonstrated in field and laboratory experiments (Silvester 1989, Barron et al. 2008, Wurz-burger et al. 2012, Jean et al. 2013, Reed et al. 2013); the extent, however, of Mo and P limitation in symbiotic N-fixing lichens is not known.
Epiphytic lichens accumulate nutrients passively through a combination of atmospheric deposition (Stewart et al. 1995, Clark et al. 1998) and canopy throughfall (Goward andArsenault 2000, Campbell et al. 2010).Their inability to directly access soil reservoirs of rock-derived nutrients, such as Mo and P, results in the potential for limitation by these elements.Indeed, several studies have demonstrated P limitation in N-fixing lichens on the oldest part of the Hawaiian archipelago and in the Pacific Northwest of North America (Kurina and Vitousek 1999, Benner et al. 2007, Benner and Vitousek 2007, McCune and Caldwell 2009).Others, however, have found no response to P-fertilization (Johansson et al. 2011), or no consistent differences in lichen N-fixation rates (Bidussi et al. 2013) or tripartite cyanolichen abundance (Campbell et al. 2010) across P gradients.When P-fertilized L. pulmonaria were examined across an elevation gradient in British Columbia, direct and indirect light exposure and bark pH, but not P treatment, affected relative growth rates (Bidussi et al. 2013).Phosphorus fertilization of lichen communities in the arctic tundra increased N-fixation rates in the cyanolichens Peltigera aphthosa and P. polydactyla but did not correspond to a change in abundance of these lichens at the plot-level (Weiss et al. 2005).
While the influence of P on lichen growth has received some attention, the role of Mo and V in cyanolichen nutrient requirements is largely unexplored.Fertilization with Mo has been attempted in only one study of cyanolichens (Horstmann et al. 1982), and to our knowledge, co-limitation of cyanolichens by V, Mo, and P has not been studied.The discovery that many commercial P-fertilizers contain Mo as a contaminant (McBride andSpiers 2001, Molina et al. 2009) has raised concern that the positive growth and fixation responses of early P-fertilization studies were actually the result of increased Mo (Barron et al. 2008).The only existing controlled study of Mo effects on cyanolichens found that the addition of 1 lg/g Mo increased nitrogenase activity, as measured by acetylene reduction, by 180% in L. pulmonaria and by 50% in Lobaria oregana (Horstmann et al. 1982).No one has demonstrated Mo limitation of cyanolichen growth and N fixation under natural conditions.
To better understand the effects of these nutrients and their interactions on lichen growth rates, we tested for P, Mo, and V limitation in the N-fixing cyanolichen L. pulmonaria with a oneyear field experiment.We chose our study species, Lobaria pulmonaria (henceforth ''Lobaria''), because it is ecologically significant, widely distributed, well studied, and easily transplanted.Not only is it of conservation concern as an old-growth associate in Europe, but also, it is sensitive to atmospheric deposition of sulfur and N (e.g., Geiser and Neitlich 2007).By determining fertilization effects on concentrations of Mo, V, and P in the lichen thallus, as well as thallus mass, natural abundance d 15 N, total %N, and change in thallus N content over the 1-year experiment, we aimed to determine if Mo, V, or P limited or co-limited growth and N fixation in Lobaria.We also fertilized the non-N fixing green algal lichen Usnea longissima Ach.(henceforth ''Usnea'') to distinguish nutrient limitation patterns between fixing and non-fixing taxa.Tissue nutrient values were used to evaluate whether one-time applications of Mo, V, and P were retained by the lichen thallus for one year in the field.Natural abundance d 15 N values were used to assess the degree to which N in Lobaria originated from fixed atmospheric N 2 vs. ambient N in the environment.The few existing cyanolichen Mo limitation studies have applied micronutrients in the field or lab and measured acetylene reduction rates (a proxy for N-fixation) within 13 hours of treatment.This study is the first to document retention of a one-time application Mo and V over the course of a year-long field transplant.
Analysis of Lobaria growth rates in the current experiment revealed much faster growth than expected compared to a previous experiment from a nearby site.To reconcile this difference between studies, we compared climate and N deposition at the two study sites, and also analyzed archived samples from the prior experiment (2009) for P, Mo, V and N, to compare to our 2010-2011 data.These comparisons allow us to determine which factor(s) may explain wide differences in growth rates and nutrient responses in the two experiments.

Specimen collection, fertilization, and transplant
Lobaria samples were collected at the experimental site where a Quercus-Pseudotsuga woodland transitions to grassland in the Coast Range foothills of Douglas County, Oregon (43.082268N, 123.522078W), southwest of Roseburg, near the town of Tenmile.Because Usnea was not present at the Tenmile site, this species was collected from an area with a similar climate in Polk County, Oregon (44.786888 N, 123.413648 W), in openings along a low-traffic country road.The road traverses foothills of the Coast Range, slightly above the Willamette Valley through forest dominated by Pseudotsuga menziesii with patches of hardwoods such as Acer macrophyllum, Crataegus spp.and Cornus nuttallii.Although we were unable to collect Usnea at Tenmile, the experimental site is within the natural range of habitat conditions for this species.Usnea transplants from an earlier study are thriving near the Tenmile site, even ten years after the initial transplant, and natural populations exist within 1 km of the site.We treated detached lobes of Lobaria and detached Usnea segments of uniform length with Mo, V, and P, then transplanted these specimens to the study site in Douglas County, Oregon, U.S.A. (43.082268 N, 123.522078 W), where they grew for one year with no further treatment.
Terminal lobes of similar size (3-6 cm long) were clipped from healthy Lobaria thalli lacking apothecia prior to treatment, resulting in air-dry pre-transplant weights of 0.1-0.4g.Usnea strands were clipped to lengths of 20 cm prior to treatment, with weights of 0.1-0.2g.All airdried specimens were weighed before treatment to obtain initial biomass, using additional sacrificial samples to correct to an oven-dry basis following McCune et al. (1996).Both species were affixed to monofilament loops with clear silicone sealant and small plastic tags (Fig. 1).
Because commercial P fertilizer has been shown to contain Mo as a trace contaminant (McBride and Spiers 2001), we tested Mo and V concentrations in three replicates each of three types of P fertilizer; commercial grade triple superphosphate, reagent grade KH 2 PO 4, and TraceSelect metal-free K 2 HPO 4 before selecting a P salt for fertilization.Our direct measurements found that triple superphosphate contained on average 9.2 lg/g Mo, while reagent grade and TraceSelect K 2 HPO 4 contained 0.21 and 0.13 lg Mo/g, respectively.On average, triple superphosphate contained 597 lg/g V, while reagent grade and TraceSelect K 2 HPO 4 contained 0.26 lg/g and 0.21 lg/V g, respectively.Thus, the TraceSelect K 2 HPO 4 used in this study contributed 0.0038 lM Mo and 0.012 lM V contamination in the P nutrient solutions.These trace contamination levels in the phosphorus treatments are 541 and 1037 times lower than our deliberate Mo and V additions respectively.
Treatments of Mo, V, P, Mo þ P or V þ P were applied to 20 randomly selected samples of each lichen species.Concentrations of Mo and V were selected in proportion to naturally occurring ratios of Mo:P or V:P reported for each lichen species (see ''Clean Site Thresholds'' of Geiser 1996).All treatments were applied by immersing the dry lichen lobes for 20 minutes in a nutrient solution made with de-ionized water.Phosphorous was applied as a 0.016 M solution of TraceSelect K 2 HPO 4 .Molybdenum was applied as a 2.1 lM solution of ACS Reagent Grade Na 2 MoO 4 2H 2 O. Vanadium was applied as 12.1 lM solution of ACS Reagent Grade Na 3 VO 4 .Combination treatments were produced by immersing lichens in a solution consisting of 2.1 lM Mo or 12.1 lM V in 0.016 M K 2 HPO 4 .Thirty control specimens of Usnea and 32 control specimens of Lobaria were submerged in deionized water in the same manner, for a total of 262 specimens (McCune et al. 1996; Fig. 1A).
After fertilization, samples were air-dried indoors overnight and transplanted into the field the following day.Three additional treated, airdried specimens each of Mo, P, V and control Lobaria and three specimens each of Mo, P, and control Usnea were sacrificed and analyzed to determine if the nutrient additions had been retained.Specimens were transplanted outdoors on October 15, 2011 to the Tenmile site described above, where they were hung in a random configuration on wooden racks (Fig. 1B).Transplants racks were constructed from a central post and several outward radiating 2.5 cm diameter dowels positioned in all compass directions from the posts (Fig. 1C).Transplant racks were fenced to discourage herbivory by deer.Lichen transplants were recovered one year later, on October 13, 2012.Of the original 262 specimens, four Lobaria (2 Mo þ P, 1 V þ P, and 1 P-treated) were either missing entirely or visibly fragmented.The remaining 258 lichen transplants were retrieved in dry weather and transported to Oregon State University on the same day.Specimens were airdried overnight and re-weighed.

Growth rates
Annual percent growth was calculated based on initial (M 0 ) and post-transplant (M 1 ) biomass measurements as G ¼ 100 3 (M 1 À M 0 )/M 0 .Because lichen masses are highly sensitive to changes in humidity, moisture correction data were obtained during weighing by reweighing the same two specimens of each species after every tenth specimen.This allowed us to correct for changes in mass due to changes in humidity during the weighing process.After subtracting out the mass of the inert apparatus and correcting for ambient moisture, pre-and post-transplant weights were adjusted to a dry-mass basis using three oven-dried sacrificial lichens of each species (McCune et al. 1996).Chemical analyses were performed on subsets of the replicates used in the field transplant, after growth rates had been determined.

Total nitrogen and d 15 N analysis
Twenty-four control and 19 P-treated Lobaria and 12 control and 9 P-treated Usnea transplants from Tenmile were analyzed for stable isotope d 15 N and percent nitrogen (%N) at the Colorado Plateau Stable Isotope Laboratory, Northern Arizona University.Two specimens of Mo, V, V þ P, and Mo þ P treatments of each species were also measured for d 15 N and %N.Change in N content (DN content) was calculated using the measured final %N for each specimen and the average initial %N of archived untreated specimens from the 2011 field collection, along with the final and initial masses:

Elemental analysis for Mo, V, and P
A subset of treated lichens from Tenmile (2-3 per treatment, per species for a total of 32) was analyzed for elemental concentrations after 1 year in the field (see Appendix: Table A1 for sample sizes).We also analyzed 3 P-treated and 5 control Lobaria archived from the 2009 Wren study, and 8 Lobaria collected from Wren in 2012.The small subsample size provides low power for detecting statistically significant differences, but does provide some indication of the degree to which nutrients provided by the initial treatments persisted through a year in the field.
In preparation for elemental analyses, all transplanted lichens were oven-dried overnight at 608C and ground to a fine powder using a Retsch PM 100 agate ball-mill to avoid metal contamination.Before acquiring the agate ball mill, we also tested two other commonly-used grinding methods for potential Mo and V contamination; a Wiley mill and a steel bar roller mill.Mo concentrations in roller mill ground lichens were between 25 and 390 times higher than those ground by Wiley mill, which in turn had Mo concentrations double those of unground lichens (Appendix: Table A2).Both roller mill and Wiley mill ground Lobaria had more Mo contamination than did ground Usnea, most likely because Lobaria required longer grinding times.Neither Wiley nor roller mill produced V contamination for Usnea.The Wiley mill did not introduce V contamination for Lobaria, but Lobaria ground by roller mill had V concentrations 1.7 times higher than that of intact specimens (Appendix: Table A2).We concluded it was necessary to use a grinder without steel components for studies of Mo and V in plant material.
All specimens were pre-digested at room temperature in microwave vessels for 12 hours prior to a heated, pressurized digestion at 2408C and 45 bar.Approximately 200 mg of lichen powder was microwave digested via Anton Paar Multiwave 3000 in 2 mL 30% Aristar Ultra-pure H 2 O 2 and 6mL 70% Aristar Ultra-pure HNO 3 (Moreira et al. 2005).Molybdenum and V concentrations were measured by Inductively Coupled Plasma Mass Spectrometry (ICP-MS), and P concentrations were measured by Inductively Coupled Plasma Optical Emission Spectrometry ( ICP-OES) at the W.M. Keck Collaboratory for Plasma Spectrometry at Oregon State University.Analyses were performed after digests were diluted 25-fold with 1% quartz distilled HNO 3 .Stock solutions of Mo, V, and P were used as calibration standards.Two separate plant standards were measured during each analysis for quality control; one US.National Institute of Standards and Technology (NIST) plant standard reference material (SRM1515 apple leaf ) and one International Atomic Energy Agency (IAEA) plant standard (IAEA-336 lichen).Analytical uncertainties for P, V, and Mo concentrations were approximately 5%.

Climate and nitrogen deposition
Both the Tenmile and Wren field sites are located in the Coast Range of Oregon and share a semi-Mediterranean climate defined by cool, wet winters and warm, dry summers.We used gridded weather interpolations (PRISM Climate Group 2014) for the 12 months corresponding to each transplant experiment to determine if temperature, precipitation, or dew point might relate to different growth responses and patterns of nutrient limitation.
Because N deposition rates can interact with P cycling in lichens (Hogan et al. 2010, Sparrius et al. 2013), we estimated total N deposition at the sites of both experiments.In particular, we expected that elevated N deposition might suppress N-fixation and growth at the Wren field site, given its closer proximity to developed areas of the Willamette Valley.To estimate N deposition we applied the Forest Inventory and Analysis (FIA) lichen indicator method (Geiser andNeitlich 2007, Jovan 2008).Macrolichens were inventoried in one 0.4 ha plot centered on each transplant site and mean annual deposition of N was estimated from the resulting list of eutrophic and oligotrophic lichens, based on the tables and regression models summarized in McCune and Geiser (2009).

Statistical analysis
Treatment effects on growth rate, %N, DN content, and d 15 N were determined using ANOVA in R version 3.1.0(R Core Team 2014).Because growth rates depended on initial mass for both species, we removed the effects of initial mass prior to testing for differences among treatments.Combining data from all treatments, growth rate (G) had a strong negative exponential relationship to initial mass (M 0 ), a trend not seen by McCune and Caldwell (2009).Thus, in the current study, thalli that were smaller prior to transplanting grew faster than thalli that were initially larger.This has been reported previously for Lobaria (Denison 1979) and other lichen species (Rhoades 1977, Benedict andNash 1990).Initial and final masses of the lichen thalli were log transformed to linearize and to remove heteroscedasticity in that relationship.We then regressed final mass (log M 1 ) on initial mass (log M 0 ), which eliminated the curvilinear relationship between G and M 0 (Lobaria, r 2 ¼ 0.87; Usnea, r 2 ¼ 0.95).The residuals of those regressions were then used as adjusted growth rates.Adjusted growth rate can be thought of as the difference between observed and expected values of log(growth), based on initial thallus size.For example, an adjusted growth rate that is a difference of 0.05 corresponds to a growth rate that is 1.12 times (or 12%) greater than expected based on initial size alone.We tested for treatment differences with simple one-way ANOVA using the adjusted growth rates.Although we made no formal tests of normality and homogeneity of variances, adjusted growth rates were similarly variable within groups; boxplots of other response variables showed no extreme outliers or skewness.
When comparing Mo, V, and P concentrations, %N, and d 15 N among treatments and across the two transplant locations, small and variable sample sizes (2-3 lichens per treatment) do not allow for robust statistical analyses.For several statistical analyses, lichens receiving a given treatment have been pooled together (as indicated in figures and figure captions).This increases sample size when comparing retention of an individual nutrient.For example, when assessing the retention of Mo in Mo-fertilized Lobaria, we considered Mo and Mo þ P Lobaria together as ''Mo-treated,'' while C, V, P, and V þ P Lobaria were ''non-Mo-treated.''We therefore evaluated these data using one-way ANOVA.

Efficacy and retention of nutrient treatments
Immediately following nutrient application, P treatment increased P concentrations in Lobaria from medians of 2226 to 3002 lg/g (Fig. 2).Starting concentrations of P were much lower in Usnea than in Lobaria, and P treatment increased median P concentrations in Usnea from 458 to 1045 lg/g.Molybdenum treatment increased Mo concentrations from medians of 59 to 1252 ng/g in Lobaria and from 68 to 362 ng/g in Usnea.
Vanadium treatment increased V concentrations in Lobaria from a median of 736-6662 ng/g.Postfertilization V concentrations were not measured in treated Usnea.To ensure that the treatment handling did not introduce Mo, V, or P to control Lobaria, we also measured nutrient concentrations in natural samples that were not handled for the experiment (Fig. 2).We found no difference in Mo, V, or P concentrations in control and natural Lobaria ( p .0.35).
After 1 year in the field, P, Mo, and V concentrations increased further in treated and control specimens of both species (Fig. 2), rather than being diluted as the thalli grew ( p , 0.001).Both P-treated and non P-treated Lobaria exhibited P gains of ;8% on average compared to pretransplant concentrations, while both Mo and non-Mo treated Lobaria exhibited Mo gains of ;70% on average.Lobaria that were not treated with V exhibited average V increases of ;50% compared to pre-transplant concentrations, while V-treated Lobaria exhibited average V increases of ;30%.P-treated Lobaria remained higher in P (median ¼ 3347 lg P/g) than non-P-treated Lobaria (median ¼ 2303 lg P/g; Fig. 2).On average, Ptreated Lobaria contained 33% more P than their non-treated counterparts.Similarly, P concentra-tions in P-treated Usnea were on average 131% higher than non-P treated Usnea (medians of 1431 lg P/g and 483 lg P/g, respectively).After one year, Mo-treated Lobaria contained 17 times more Mo on average (median ¼ 4663 ng Mo/g) than non-Mo treated Lobaria (202 ng Mo/g).Motreated Usnea also remained higher (median ¼ 698 ng Mo/g) than non-Mo-treated Usnea (320 ng Mo/g).Mo-treated Usnea contained nearly 3 times more Mo on average than their non-treated counterparts.Finally, V-treated Lobaria displayed higher V after one year in the field (median ¼ 9838 ng V/g) than non-V-treated Lobaria (1292 ng V/g).Vanadium treated Lobaria and Usnea contained an average of five and two times more V respectively than non-treated counterparts of the same species.

Annual average growth rates
Unadjusted growth rates for all lichens were high regardless of treatment, averaging 83% for Lobaria and 108% for Usnea over the course of the 1-year experiment.All 130 Usnea and 32 of 128 Lobaria at least doubled in size over the course of the year.These high growth rates were evaluated with a repeat experiment using new transplants collected at Tenmile without nutrient treatment, but on the same racks; these showed considerably lower growth rates during October 2013-October 2014 than at the same site 2 years earlier (Lobaria mean growth rate: 20%, n ¼ 24; Usnea 29%, n ¼ 18).
Adjusted growth rates did not differ significantly for control, P, Mo, Mo þ P, and V treatments of Lobaria ( p-value for all comparisons .0.07).Neither Mo þ P nor V þ P treated Lobaria grew faster than the controls (Fig. 3).VþP treated Lobaria grew less than control Lobaria ( p ¼ 0.005), but the growth rate of that treatment was only 3% less than the overall average growth rate (Fig. 3).No treatment effect on adjusted growth rate was observed for Usnea ( p ¼ 0.89; Fig. 3).
Total thallus N (%) in Lobaria was generally similar across all treatments and sites.In Lobaria at Wren however, %N averaged slightly (0.2%) higher in controls than with P treatment ( p ¼ 0.030).

Local climate and nitrogen deposition rates
Estimated average monthly dew point and average monthly maximum and minimum temperatures for Wren (1 October 2007-31 October 2008) and Tenmile (1 October 2011-31 October 2012) were similar between sites.Total precipitation for the 13 month periods was nearly the same (Wren, 1077 mm; Tenmile, 1122 mm), but some potentially important seasonal differences emerged.Previous work near the Wren site found that growth of Lobaria peaked in March and April, while no consistent growth occurred during the dry summer months and early autumn (Muir et al. 1997).For February through May, total precipitation at Wren was 257 mm in 2007-2008, while Tenmile received 511 mm during February-May of 2010-2011.
We estimated total annual N deposition for Tenmile and Wren sites at 3.6 and 3.0 kg NÁha À1 Áyr À1 , respectively, based on lichen surveys of communities containing 45 macrolichen species at Wren and 38 species at Tenmile.McCune and Geiser (2009) estimated confidence limits for this method as a 70% chance of being within 1 kg NÁha À1 Áyr À1 of the total N deposition.The two sites shared 26 species.Both sites supported a preponderance of oligotrophic (N-intolerant) and mesotrophic species, with a few eutrophic species.Thus, based on this method, atmospheric N deposition was similar between the two sites.

Naturally occurring nutrient concentrations
Control Lobaria from Tenmile averaged 60% higher P concentrations than control Lobaria from Wren (2431 lg/g versus 1474 lg/g, p ¼ 0.024; Fig. 5).Similarly, natural Lobaria from Tenmile (2043 lg/g, n ¼ 5) were higher in P than natural Lobaria from Wren (1299 lg/g, n ¼ 5; p ¼ 0.018).However, natural and control Lobaria P concentrations from Tenmile were similar to P-fertilized Lobaria from Wren (McCune and Caldwell 2009; both p .0.27) and were at least twice those of untreated Fig. 3. Adjusted growth rates of Lobaria and Usnea for each treatment.Growth residuals are equivalent to 10 y where y is the ratio of final mass to expected final mass, based on initial mass.For example, a residual value of 0.05 corresponds to a growth rate that is 1.12 times (or 12%) greater than expected based on initial size alone.Boxes represent the 25th and 75th percentiles, horizontal lines within each box represent medians, and whiskers indicate the 5th and 95th percentiles; symbols indicate outliers that fall more than 1.5 times the inter-quartile range from the median.v www.esajournals.orgspecimens from that prior study ( p ¼ 0.005).Mean concentration of P in control Lobaria from Tenmile did not differ from natural, un-transplanted archived Lobaria, which were stored in the lab ( p ¼ 0.658).Nor did mean P concentration of natural Lobaria collected from Wren in 2012 differ from controls from the 2007 field study ( p ¼ 0.508).
Molybdenum concentrations were similar between sites but higher in control than natural (untreated) Lobaria at Wren ( p ¼ 0.02).Control and natural Mo concentrations in Lobaria from Tenmile were 185 and 168 lg/g, respectively; from Wren, 207 and 70 lg/g, respectively.Vanadium concentrations in Lobaria were similar among controls and natural specimens from both sites (Tenmile controls ¼ 1537; Tenmile naturals ¼ 1230 lg/g; Wren naturals ¼ 1589 lg/g), except that V in Wren controls (2994 lg/g) was notably high ( p , 0.001).

DISCUSSION
The lack of a strong growth response for fertilized Lobaria in our experiment at Tenmile contrasts with the doubling of Lobaria growth rates with P fertilization observed previously at the Wren site by McCune and Caldwell (2009), using the same P-treatment concentrations, on the same species, in the same climate regime.Although the favorable weather and consistently high growth rates we observed at Tenmile should correspond with high P-demand, P did not limit Lobaria growth in the current experiment.We attribute the fast growth and lack of a P limitation observed in our Tenmile experiment v www.esajournals.orgto favorable weather conditions during the transplant year, and high levels of available P at the Tenmile site.Our results provide evidence that annual growth rates of Lobaria depend on both site-specific differences in background P and weather conditions in particular years.

Nutrient retention
After one year in the field, fertilization treatments were retained in the lichen thallus in both Lobaria and Usnea.Furthermore, additional Mo, V, and P accumulated over the duration of the transplant in both species.The retention and accumulation of Mo, V, and P does not explain why these nutrients failed to stimulate Lobaria growth, but it does suggest nutrient assimilation regardless of the lichen's need for nitrogenase cofactors.The passive accumulation of metals in lichen thalli is well documented in studies used to monitor pollution (e.g., Addison andPuckett 1980, Garty 2001, andothers).

Effects of climate and N deposition
We considered differences in climate and N deposition as among the leading possible explanations for the difference in results of the two experiments.Differences in N deposition could influence lichen growth directly due to cyanolichen sensitivity to ambient N, as well as indirectly via interactions between N and P (e.g., Hogan et al. 2010, Johansson et al. 2011, Sparrius et al. 2013).Although high ambient N can cause down-regulation of N-fixation, both sites had similar estimated N deposition, suggesting that deposition did not alter response to P-fertilization at one site versus the other.We also found that d 15 N of Lobaria at both Tenmile and Wren field sites was close to the atmospheric value of 0%, consistent with physiological studies indicating that Lobaria can meet much of its N demand via fixation (Antoine 2004).If we assume that Usnea records the d 15 N of ambient deposition and Lobaria fixes N 2 without fractionation, then simple end-member mixing suggests that Lobaria derives ;86% of its N from fixation at Tenmile.Because Usnea was not included in the 2009 study at Wren we are unable to create a similar mixing model for that site.Small differences in d 15 N values of Lobaria at the two sites may reflect differences in N fixation, but additional information on d 15 N of other potential N sources is needed to evaluate this possibility.In  2002, Felix et al. 2013, Felix et al. 2014).
Faster growth of Lobaria at Tenmile than Wren may be attributable to differences in growingseason precipitation.Although temperature, dew point, and total annual precipitation were comparable, the two sites differed in the distribution of annual rainfall.The scale and duration of individual rain events cannot be assessed.However, the Tenmile site received about 511 mm of precipitation during the peak growing season (February-May) of that year, about twice as much as during the corresponding period at Wren (257 mm) during that experiment.Overall, mean annual growth rates for Lobaria of all treatments from the Tenmile experiment are among the highest reported for Lobaria in Pacific Northwest of North America.The growth rates reported in this study are 20-50% higher than growth rates of P-treated Lobaria reported for the Wren site (McCune and Caldwell 2009).They are also higher than the annual growth rates two years later at the same site, and higher than growth rates reported for any other year-long study in this region (e.g., McCune et al. 1996, Muir et al. 1997, Antoine and McCune 2004).The high growth rates in the Tenmile experiment suggest that minimal environmental constraints (including nutrients) limited the expression of a fertilizer response at this site.

Environmental and physiological effects
Lobaria growth rates can be affected by light and moisture (Gauslaa et al. 2006, Gauslaa 2014).Although we did not measure incident light at either site, we assume that our artificial transplant racks were exposed to more direct light than epiphytes growing on branches.Thus, it is possible that the difference in light regime between Tenmile (where lichens were hung from racks) and Wren (where lichens were hung from tree branches) contributed to the difference in growth response between sites.Additionally, canopy cover can reduce nocturnal cooling and increase dewfall in branch microenvironments, both of which have been show to increase growth in Lobaria (Bidussi et al. 2013).However, our results show higher growth rates in Tenmile, where lichens were more exposed on transplant racks.We cannot, therefore, rule out the possibility that the overall faster growth at Tenmile than Wren was a result of differences in light environment and moisture retention between experiments.Neither of these, however, can explain the lack of a P treatment effect at Tenmile versus a strong P effect at Wren.
Recent research suggests that there is a tradeoff in Lobaria between thallus growth, measured both as dry matter gain and increased surface area, and the production of asexual reproductive structures (soralia and/or isidia; Gauslaa 2006).Although we were able to limit our transplant specimens to those lobes lacking sexual reproductive structures (apothecia), the quantity of soralia varied in our specimens.We cannot rule out the influence of soralia production on the observed growth rates, though our large sample size and random application of treatments should have served to minimize that effect.

Molybdenum retention and potential sources
The lack of a significant effect of V and Mo, whether alone or in combination with P, on Lobaria growth rates suggests that neither micronutrient was limiting at Tenmile.However, the gains in P, Mo, and V in Lobaria transplants after one-year at the Tenmile site suggest that this environment is naturally rich in those nutrients.The average concentration of Mo in non-Mofertilized Lobaria from Tenmile was 227 ng/g 1 yr after transplanting.When this concentration is distributed over the thallus of a lichen with an average surface area of 24 cm 2 and an average dry mass of 0.3-0.5 g, Tenmile lichens contained 2.8-4.7 ng Mo per cm 2 of thallus.Using the average Mo concentration in openfall rainwater measured across the Coast Range (17 pg/g; Marks 2015) and the approximately 1122 mm of precipitation that occurred from 2011-2012 at the Tenmile site, we determined that this site could have received 1.9 ng Mo per cm 2 of thallus in that year.Although cyanolichens beneath the forest canopy may not directly intercept all of the incoming rainfall, Mo concentrations of this magnitude could be an important Mo source over several years of thallus growth, in addition to an unknown amount of Mo derived from dry deposition or leached from the canopy as part of throughfall.Variations in the quantity and distribution of annual rainfall may affect not only thallus hydration, but also the amount of Mo deposition and retention.

Variations in phosphorus availability
Our data clearly show that naturally available P is more abundant at Tenmile than at Wren.Phosphorus concentrations in natural and control samples of Lobaria at Tenmile were on average 1.67 times higher than in natural and control samples from Wren.Large differences in naturally available P between sites could easily affect responses to P fertilization.P fertilization at Wren considerably boosted P in lichen thalli relative to controls, yet the fertilized P values were similar to those occurring naturally at Tenmile.Combining results from the two experiments, Lobaria grew fastest when internal P concentrations reached ;2000 lg/g.Above this concentration, additional P did not result in additional growth, suggesting that any P limitation had been overcome.It therefore appears that growth of Lobaria at Wren was P-limited, while growth at Tenmile was not.
Lichens are subjected to potential P sources that vary greatly in magnitude, including local and regional atmospheric deposition, and canopy leachate.Differences in P availability among sites may be caused by P recycling in organic matter of the dominant tree species, geomorphology, canopy cover, the presence or absence of edge effects, proximity to intensive agriculture, or P content of soil and bedrock (Walker and Syers 1976, Benner 2011, Cross and Perakis 2011, Mage and Porder 2013).More research would be necessary to determine the factors responsible for the differences in available P levels at our sites.In addition to geologic differences among sites, atmospheric inputs of P are both spatially and temporally variable in western Oregon (Fredriksen 1975); preliminary data suggest that differences in soil P between sites were not responsible for differences in lichen P contents.Analysis of lichen P concentrations may help assess whether P is locally limiting in cyanolichens and reconcile the differing results of existing cyanolichen fertilization experiments.

Fig. 1 .
Fig. 1.Transplant construction and deployment.(A) Post-treatment Lobaria transplants air drying in the lab.(B) Field-dry Lobaria and Usnea transplants attached to wooden rack with zip ties.(C) Transplant racks at the Tenmile field site, October 2012.

v
www.esajournals.orgdifference in DN content between control and Ptreated Lobaria at Tenmile ( p ¼ 0.16 ).To understand the lack of definitive growth and DN content responses among treatments, we compared d 15 N and %N in our study at Tenmile with Lobaria samples from the McCune and Caldwell (2009) study at Wren, where P treatment doubled Lobaria growth.Average d 15 N values in control and P-treated Lobaria from Tenmile were similar (Fig. 4; control ¼ À1.7%,

Fig. 2 .
Fig.2.Efficacy and retention of nutrient treatments.Concentrations of P, Mo, and V in nutrient-treated (red) and controls (gray) for Lobaria (left) and Usnea (right) at three dates.''Pre'' was after preparation of transplants but before treatment; ''Post'' was immediately after treatment with the indicated nutrient, and demonstrates the initial efficacy of the treatment; ''One Year Post'' was after one year in the field and demonstrates retention of the treatments.Sample sizes are indicated below each box.Pre-treatment data were not obtained for Usnea.Boxes represent the 25th and 75th percentiles, horizontal lines within each box represent medians, and whiskers indicate the 5th and 95th percentiles; symbols indicate outliers that fall more than 1.5 times the inter-quartile range from the median.

Fig. 4 .
Fig. 4. Thallus d 15 N for Lobaria and Usnea at Tenmile.Treatments for Lobaria include control (n ¼ 24), phosphorus (n ¼ 19), molybdenum (Mo), vanadium (V), and the combinations Mo þ P and V þ P (n ¼ 2 each).Treatments for Usnea include control (n ¼ 12) and phosphorus (n ¼ 8).Boxplots represent the 25th and 75th percentiles, horizontal lines within each box represent medians, and whiskers indicate the 5th and 95th percentiles; no outliers fall more than 1.5 times the inter-quartile range from the median.Individual data points instead of boxplots are shown for treatments with n ¼ 2.

Fig. 5 .
Fig. 5. Annual percent growth of Lobaria in relation to thallus P concentration (lg/g), combining results from two experiments in different locations: Tenmile (this paper) and Wren (McCune and Caldwell 2009).Bars indicate interquartile ranges; black dots are medians.Phosphorus treatments (red bars) were the same for both experiments.Wren specimens are from the 2009 season only and do not include 2012 data for which growth rates are not available.