Root chemistry of Douglas-fir seedlings grownunder different nitrogen and potassium regimes

Terry M. Shaw, James A. Moore, and John D. Marshall

Abstract: Root chemistry and biomass allocation of Douglas-fir (Pseudotsuga menziesiivar. glauca (Bessn.) Franco)seedlings under optimal and deficient levels of nitrogen (N) and potassium (K) were studied. Seedlings receivinghigh-N treatments were significantly larger and allocated more dry matter to their stems and less to their roots thanthose receiving the low-N treatments. The K treatments did not significantly affect total seedling biomass or root/shootratios. Root tip starch concentrations were significantly higher and root tip sugar concentrations were lower in plantsreceiving the low-N treatments. Seedlings receiving the high-N, low-K treatment had significantly lower concentrationsof phenolics and tannins and lower ratios of these compounds to sugars in the root tips than seedlings receiving thehigh-K treatments. Samples taken from two locations on the root system show that concentrations of phenolics,tannins, sugars, and starches were substantially higher in the root collar than in the root tips. Because of lower withintissue variation, we recommend sampling at root tips to better detect treatment differences. This study shows that Nlevels affect starch concentrations in the roots, while K levels affect root phenolic and tannin concentrations. Possiblerelationships between low root phenolic and tannin concentrations and lessened resistance of Douglas-fir to root diseaseare discussed.

Résumé: La chimie des racines et l’allocation de la biomasse ont été étudiées chez des semis de Douglas taxifolié(Pseudotsuga menziesiivar. glauca (Bessn.) Franco) soumis à des niveaux faible ou optimal d’azote (N) et depotassium (K). Les semis soumis au niveau élevé de N avaient une dimension significativement plus forte et allouaientplus de matière sèche vers la tige que vers les racines comparativement aux semis soumis au niveau faible de N. Lestraitements avec K n’ont pas significativement affecté la biomasse totale des semis ni le ratio racines/tige. Laconcentration en amidon était significativement plus élevée et la concentration en sucre plus faible dans l’apexracinaire des plants soumis au faible niveau de N. Les plants soumis aux niveaux élevé de N et faible de K avaientdes concentrations significativement plus faibles de composés phénoliques et de tannins et des ratios plus faibles de cescomposés par rapport aux sucres, dans l’apex racinaire, que les semis soumis au niveau élevé de K. Des échantillonsprélevés à deux endroits du système racinaire montrent que les concentrations de composés phénoliques, de tannins, desucres et d’amidon étaient substantiellement plus élevées au collet qu’à l’apex racinaire. Étant donné la plus faiblevariation à l’intérieur de ces tissus, nous recommandons d’échantillonner à l’apex racinaire pour mieux détecter lesdifférences entre les traitements. Cette étude montre que le niveau de N affecte la concentration d’amidon dans lesracines tandis que le niveau de K affecte les concentrations de composés phénoliques et de tannins dans les racines.On discute de la relation qui pourrait exister entre de faibles concentrations de composés phénoliques et de tanninsdans les racines et une moins grande résistance du Douglas taxifolié aux maladies des racines.

[Traduit par la rédaction] Shaw et al. 1573

Concentrations of storage and secondary compounds inplant tissue depend to a considerable extent on the environ-ment in which plants grow (Waring et al. 1985; Huber andArny 1985). In particular, mineral nutrition influences con-centrations of secondary carbon compounds such as sugars,starches, phenolics, and tannins (Bryant et al. 1983; Entry etal. 1991a; Moore et al. 1994). The concentrations of thesecompounds and the balance among them help to determinethe resistance of plants to herbivores and pathogens (Wargo

1972; Garraway 1975; Ostrofsky and Shigo 1984; Larsson etal. 1986; Mwangi et al.1990; Dudt and Shure 1994). There-fore, the levels of available nutrients such as N and K mayinfluence the ability of plants to resist disease. Douglas-fir(Pseudotsuga menziesiivar. glauca(Bessn.) Franco) was se-lected for this experiment because it is commonly managedin the forests of the Inland Northwest and has been the sub-ject of many previous fertilization field experiments (Heathand Chappell 1989; Shafii et al. 1989, 1990; Moore et al.1991). Based on foliar analysis and growth response resultsfrom these experiments, nitrogen and potassium are com-monly deficient (Mika and Moore 1991; Moore et al. 1994).Thus, N and K were chosen for experimentation in thisgreenhouse study.

One control over the ability of a tree to resist stress is theavailability of carbohydrate reserves (Waring and Schlesinger1985). However, competition for photosynthate may reducethe levels of carbohydrate reserves available for mobilizationin various tissues. Resource availability may also influence

Can. J. For. Res.28: 1566–1573 (1998) © 1998 NRC Canada

1566

Received February 11, 1998. Accepted July 31, 1998.

T.M. Shaw, J.A. Moore,1 and J.D. Marshall. Department ofForest Resources, University of Idaho, Moscow, ID 83844,U.S.A.

1Author to whom all correspondence should be addressed.e-mail: jamoore@uidaho.edu

production of secondary compounds like phenols and tan-nins (Mooney 1972; Bazzaz et al. 1987).

Changes in plant tissue composition result from competi-tion for photosynthate among compounds. These changescan be predicted from the carbon–nutrient balance (CNB)hypothesis (Bryant et al. 1983), which suggests that carbon-rich compounds such as phenolics and tannins are producedin relatively greater amounts when photosynthate is moreavailable and mineral nutrients are less available. The CNBhypothesis is an extension of the earlier growth–differentiationbalance hypothesis (Loomis 1932; Herms and Mattson1992), which suggests that growth and differentiation com-pete for photosynthate. Both hypotheses predict increases inthe concentration of carbon-rich compounds when nutrientsare less available. We therefore examined correlations be-tween plant biomass, phenolics, tannins, sugars, and starchconcentrations, as well as ratios of compounds occurring inthe roots.

To better estimate and understand the relationship be-tween plant nutrition, root chemistry, and plant susceptibilityto disease, the effects of sampling location on root chemistrymust also be understood. Wargo et al. (1972) found that glu-cose and fructose concentrations in sugar maple (Acer sac-charum Marsh.) were higher in the outermost root woodthan the inner root bark. In another study with sugar maple,Parker and Houston (1971) found that levels of sugars werehigher in root bark than in root collar bark. Likewise, levelsof storage and defensive compounds have been shown tovary considerably along the gradient of stem and root bark(Kelsey and Harmon 1989).

The principal objective of this study was to determine theeffects of varied N and K regimes on the chemical composi-tion of Douglas-fir roots. A second objective was to evaluatetwo sampling locations by comparing variations in theirchemical composition.

TreatmentsOne hundred and four 1-year-old, containerized Douglas-fir

(Pseudotsuga menziesiivar. glauca (Bessn.) Franco) seedlingswere planted in 2.9-L plastic containers filled with medium-gradesilica sand. Seedlings were grown at the University of Idaho ForestNursery in a shadehouse covered by a clear, corrugated fiberglass

roof from June to December for three growing seasons (1990,1991, and 1992). To obtain wider representation of the Douglas-firgene pool, seedlings from two northern Idaho Douglas-fir seedsources collected from different locations and elevations were dis-tributed equally by treatment and block. Seedlings were randomlyassigned to four different N and K treatments within two blocks.The solution used to supply the Douglas-fir seedlings with nutri-ents was adapted from Ingestad and Lund (1979) and is considerednutritionally optimal. A stock solution was formulated with levelsfor the macronutrients as follows: N, 100; K, 65; P, 13.8; Ca, 7;Mg, 8.5; and S, 15 mg/L. The micronutrient concentrations were asfollows: Fe, 700; Mn, 400; B, 200; Cu, 30; Zn, 30; Cl, 30; Mo, 7;and Na, 3 ppm. The nutrient stock solution was modified to meetthe treatment regimes shown in Table 1. Nutrient solutions weregiven to plants by adding them to well water through an injectionsystem at a ratio of 1:100. The average pH of the solution for alltreatments was 7.1. Seedlings were irrigated as needed, and500 mL of nutrient solution was applied every 4 days throughoutthe growing season. Irrigation was reduced in late September ofeach year to allow the onset of dormancy. Periodic foliar samplingwas used to adjust treatments so that the N and K foliar concentra-tions were similar to treatment ranges observed from Douglas-firfield fertilization studies. Mika and Moore (1991) observedDouglas-fir foliar N concentrations as low as 1.0% on untreatedfield sites and as high as 1.87% on N-treated sites. In the samestudy, foliar K concentrations ranged between 0.6% on untreatedsites and 0.9% on K-fertilized sites. These results from field exper-iments, along with the results from van den Driessche (1979) andWebster and Dobkowski (1983), helped determine the target N andK concentrations in our experiment. The target concentrations inour study were as follows for N: high-N treatments, 1.9%; low-Ntreatments, 1.0%; and for K: high-K treatments, 0.8%; low-K treat-ments, 0.6%. We felt that, even if the target concentrations werenot completely achieved, an acceptable range of nutrient levelswould be produced by the experiment. Each treatment consisted ofa 3-year nutrient regime periodically adjusted to attain the targetnutrient concentrations. Early during the second growing season itbecame apparent that seedlings grown in the low-N treatments re-quired more N based on very low (<0.8%) foliar N concentrationsand visual deficiency symptoms. Thus, N was increased from 10 to25% in the treatment solution concentration during the secondgrowing season (Table 1). By the third growing season, N concen-trations had increased sufficiently to allow the low N treatments’solution concentration to again be set at 10%. To prevent nutrientleaching from the foliage, irrigation water and solution was applieddirectly to the soil through 3.2-mm (inside diameter) polyethylenedrip tubing with Roberts 180E medium flow Spot-Spitters®. Toavoid low winter temperatures, seedlings were transferred andstored at around 0°C. Daylength followed seasonal variations forMoscow, Idaho. (46°44´N).

Analysis of the mineral composition of leaf tissuesDouglas-fir seedlings were sampled in October after 3 years of

growth under different N and K treatments. Seedlings and the sandin which they were grown were carefully removed from the con-tainers, sand was gently removed from the root system, then theseedlings were put into plastic bags and stored in coolers to slowplant metabolism. Bulk foliar samples were obtained from 26 seed-lings per treatment for nutrient analysis. Seedling foliage wasstored at 1EC for up to 48 h while awaiting laboratory analysis. Inthe laboratory, roots were separated from the shoots and both sam-ples were oven-dried at 70°C for 24 h. Afterward, needles werestripped from the stem and continued to dry at 70°C for an addi-tional 24 h. A sample of 2 g, representative of all the needles, wastaken from each seedling and ground in a Wiley mill in preparationfor chemical analysis. Foliar N concentration was determined usinga standard micro-Kjeldahl procedure (Bremner and Mulvaney

© 1998 NRC Canada

Shaw et al. 1567

Years Treatment Nitrogen Potassium

1990 and 1992 nk 10 10nK 10 100Nk 100 10NK 100 100

1991 nk 25 10nK 25 100Nk 100 10NK 100 100

Note: Values are percentages of solution concentrations developed byIngestad and Lund (1979).

Table 1. Nutrient treatments under which Douglas-fir seedlingswere grown: low nitrogen and low potassium (nk), low nitrogenand high potassium (nK), high nitrogen and low potassium (Nk),and high nitrogen and high potassium (NK).

1982). Phosphorus, K, Ca, Mg, Mn, Fe, Cu, and Zn concentrationswere determined by inductively coupled plasma (ICP) emission(Thermo Jarrell Ash Corp. 1988).

Quantification of carbohydrates, phenolics, and tanninsin the root tissues

To ensure sufficient material for laboratory analysis of all treat-ments, 26 root collar bark and 26 root tip composite samples (fourseedlings per composite, two from each block and seed source)were collected from the same seedlings that were used in the foliarnutrient analysis for subsequent analysis of starch, soluble sugar,total phenols and protein-precipitable tannins. Only nonmycor-rhizal long lateral living root tips ranging in diameter between 1and 3 mm were included in the samples. Roots were temporarilystored for several hours in coolers while waiting for transport fromthe nursery, then kept at –40°C until analyzed. For the chemicalanalysis, all samples (bark and root tips) were put in liquid nitro-gen overnight, root collar bark was removed to the phloem, thenthe living tissues of the inner bark were removed. Samples werethen ground to powder in a mortar. Total phenols were determinedfrom samples after extracting with aqueous acetone (80%), addingFolin–Ciocalteu’s Reagent, and then measuring absorbances at 700and 735 nm (Julkunen-Tiito 1985). Samples were analyzed for to-tal soluble starch through an ethanol and perchloric acid method(Hansen and Moller 1975) and glucose was determined by addinganthrone solution for absorbance determination at 630 nm (Hansenand Moller 1975). Concentrations of glucose were measured usinga standard curve established with a glucose standard (Hansen andMoller 1975). The Hansen and Moller method overestimates starchlevels because carbohydrates other than starch are extracted duringthe process (Marshall 1986; Rose et al. 1991). Perchlorate-extractable carbohydrates were, therefore, corrected to yield starchconcentrations and are expressed in their corrected form through-out this paper. Tannin levels were measured after extracting with80% aqueous acetone loaded into an agarose plate containing bo-vine serum albumin, diffusion rings were measured and tannin aciddetermined (Hagerman 1987). Analysis of starch, sugar (glucose),total phenols, and protein-precipitable tannins analysis were per-formed by the Institute of Biological Chemistry, Washington StateUniversity in Pullman, Wash.

Statistical analysisFrom a population of 104 seedlings, the treatment effects on fo-

liar nutrient concentrations and seedling biomass were estimatedusing analysis of variance for a 2 (blocks) × 2 (seed sources) × 4(treatments) randomized complete block design; thus, 26 seedlingsper treatment were available from each treatment. For root chemis-try analyses, six root collar bark or root tip composite samples

(composed of an equal number of seedlings from each block andseed source) per treatment were used; therefore, block and seedsource effects could not be estimated in the statistical models. Oneroot collar bark sample from the high-N, low-K (Nk) treatmentwas dropped from the analysis since starch, sugar, tannin, and phe-nolic concentrations were all more than 2.5 SDs from the treatmentmeans. Analysis of variance (PROC GLM) and differences be-tween means by treatment for foliar nutrient and root chemistrydata were determined by using the least-squares means procedureof the Statistical Analysis System (SAS Institute Inc. 1985). A de-tailed description of the ANOVA models with the degrees of free-dom associated with each term is provided in Table 2. Correlationsbetween plant biomass (g/seedling), sugar, starch, phenols, andtannin concentration by root sampling locations were analyzed us-ing PROC REG of SAS Institute Inc. (1985).

Mineral composition of leaf tissueNeither block, seed source, nor their interactions with

treatment explained significant variation in the mineral com-position of needle tissue. Foliar N concentrations were 71%higher in the seedlings receiving the high-N treatments thanin seedlings receiving low-N treatments (Table 3). Plants re-ceiving the Nk treatment had significantly lower foliar Klevels (20%) than those of seedlings receiving the high-N,high-K (NK) treatment. In addition, seedlings receiving thelow-K treatments showed K-deficiency symptoms, i.e.,chlorosis and necroses along the leaf margins. In our study,significant differences in foliar K/N ratios were observedonly between the high-N and low-N treatments (Table 3).

Seedling biomassSeedling biomass was not significantly affected by block

or seed source or their interactions with treatment. Total drybiomass after three growing seasons was significantlygreater for those seedlings receiving the high-N treatments.Seedlings grown under the low-N regime allocated a signifi-cantly larger proportion of their carbon to roots, about 50%,compared with seedlings grown with higher N supplies(Table 3). The level of K supplied in the growth regime hadno significant effect on either seedling biomass or root/shootratios (Table 3).

Carbohydrates, phenolics, and tannins in the roottissues

Individual significance levels comparing sampling loca-tions by treatment for sugar, starch, phenolics, and protein-precipitable tannins are shown in Table 4. Concentrations forthe four carbon-based compounds were all significantlyhigher at the root collar than in the root tips (Fig. 1). Solublesugar and tannin concentrations were two to three timeshigher in the root collar area, whereas concentrations ofphenolics were up to six times higher. These results were es-pecially pronounced in the Nk treatments, with phenolic andtannin levels six and seven times higher in the root collarthan in the root tips, respectively (Fig. 1). Overall, plants re-ceiving the low-K treatments had greater differences be-tween the root tip and root collar for phenolic and tanninconcentrations.

Treatment significance levels by root sampling locationfor concentration of each carbon-based compound and

© 1998 NRC Canada

1568 Can. J. For. Res. Vol. 28, 1998

df

Source A B

Block 1Seed source 1Treatment 3 3Block × treatment 3Seed source × treatment 3Error 94 22Total 105 25

Table 2. Degrees of freedom for analysis ofvariance for mineral composition of needletissue and seedling biomass (column A), andcarbohydrates, phenolics, and tannins in roottissue (column B).

phenolic/sugar ratios are provided in Table 5. Root tipchemistry was clearly affected by the treatments, except forsugar concentrations (Fig. 2a). Starch concentrations weresignificantly lower with high-N treatment levels but wereunaffected by the level of K supplied (Fig. 2b). Phenolic andtannin concentrations were significantly higher for thehigh-K treatments compared with the Nk treatment but wereunaffected by the N level supplied for a given K level(Figs. 2c and 2d). The Nk treatment produced significantlylower root tip phenolic/sugar ratios than either of the twotreatments with higher levels of K supplied (Fig. 3). Root tipphenolic and tannin concentrations were strongly affected bythe K treatments, with significant differences between theNK treatment and both low-K treatments.

Sugar, starch, phenol, and tannin concentrations were sub-stantially higher in the root collar than in the root tip, andthe results at the root collar showed similar trends by treat-ment. However, the statistical significance of the treatmentswas not as consistent at the root collar compared with theroot tips; for example, phenolic concentrations were signifi-

cantly different for the low-N, low-K (nk) versus Nk andlow-N, high-K (nK) versus NK treatments at the root collarbut not at the root tips (Table 5). Concentrations from thetwo root sampling locations were significantly correlated,having coefficients of determination of 0.72 for sugar, 0.62for starch, 0.86 for phenol, and 0.79 for tannin.

© 1998 NRC Canada

Shaw et al. 1569

TreatmentShoot drymass (g)

Root drymass (g)

Root/shootratio (%)

Nitrogen(%)

Potassium(%)

Potassium/nitrogen

nk 9.9a 9.4a 95a 1.06a 0.68a 0.67anK 9.4a 8.4a 89a 1.13a 0.71a 0.67aNk 37.7b 24.5b 65b 1.85b 0.59b* 0.33bNK 34.8b 20.3b 58b 1.90b 0.74a 0.40b

Note: Treatment abbreviations are given in Table 1. Within each column, values followed by the same letter are similar atP ≤0.01, except for the value marked with an asterisk, which is significant atP ≤ 0.05.

Table 3. Root and shoot dry masses and foliar nitrogen and potassium concentrations collected in October afterthree growing seasons from Douglas-fir seedlings.

VariableTreatmentcontrasts P*

Sugar nk <0.001nK <0.001Nk <0.001NK <0.001

Starch nk <0.001nK <0.006Nk <0.001NK <0.007

Phenolics nk <0.001nK <0.001Nk <0.001NK <0.001

Tannins nk <0.001nK <0.001Nk <0.001NK <0.001

Note: Treatment abbreviations are given in Table 1.*Probability of t greater than the calculated value

using LSD procedure of the LSMEANS–PDIFFroutine (SAS Institute Inc. 1985).

Table 4. Root tip versus root collar contrastsby treatment for sugar, starch, phenolics, andtannin concentrations collected after threegrowing seasons from Douglas-fir seedlings.

Treatmentcontrasts

P*

Variable Root tip Root collar

Sugar nk vs. nK <0.694 <0.5309nk vs. Nk <0.784 <0.4684nk vs. NK <0.634 <0.6073nK vs. Nk <0.496 <0.1759nK vs. NK <0.379 <0.2498Nk vs. NK <0.833 <0.8238

Starch nk vs. nK <0.989 <0.4984nk vs. Nk <0.001 <0.0454nk vs. NK <0.002 <0.0124nK vs. Nk <0.001 <0.1736nK vs. NK <0.002 <0.0571Nk vs. NK <0.353 <0.5379

Phenolics nk vs. nK <0.155 <0.6460nk vs. Nk <0.826 <0.0492nk vs. NK <0.061 <0.0095nK vs. Nk <0.094 <0.1242nK vs. NK <0.659 <0.0276Nk vs. NK <0.032 <0.4365

Tannins nk vs. nK <0.194 <0.7465nk vs. Nk <0.064 <0.0107nk vs. NK <0.016 <0.0001nK vs. Nk <0.003 <0.0226nK vs. NK <0.232 <0.0003Nk vs. NK <0.001 <0.0700

Phenolic/sugar ratio nk vs. nK <0.194 <0.1317nk vs. Nk <0.577 <0.2273nk vs. NK <0.121 <0.0074nK vs. Nk <0.063 <0.7054nK vs. NK <0.826 <0.1982Nk vs. NK <0.034 <0.0890

Note: Treatment abbreviations are given in Table 1.*Probability of t greater than the calculated value using LSD procedure

of the LSMEANS–PDIFF routine (SAS Institute Inc. 1985).

Table 5. Treatment contrasts for root tip and root collar sugar,starch, phenolics, tannin concentrations, and phenolic to sugarconcentration ratios collected after three growing seasons fromDouglas-fir seedlings.

Significant correlations were observed between non-structural carbohydrates and carbon-based secondary com-pounds at the root collar sampling location but not at theroot tips (Table 6). As sugar concentrations increased at theroot collar, so did phenolic and tannin concentrations. How-ever, the correlations between starch and phenolic or tanninconcentrations were negative. At the root tips, there were nostatistically significant correlations between sugar or starch

and tannin or phenolic concentrations (Table 6). Seedlingbiomass (g dry matter), showed significant positive correla-tion with tannin and phenolic concentrations at the root col-lar but evidenced no significant relationship with these samecompounds at the root tips (Table 6). Total seedling biomasswas significantly (p < 0.05) correlated with sugar and starchconcentrations at both the root collar and tips. Coefficientsof determination for biomass versus sugar and starch at the

© 1998 NRC Canada

1570 Can. J. For. Res. Vol. 28, 1998

Treatment Treatment

Su

ga

r(%

)

Sta

rch

(%)

Ta

nn

ins

(%)

(b)

(c) (d)

Ph

en

oli

cs

(%)

0

2

4

6

8

10

12

14

14

20

18

16

14

12

12

10

8

6

4

2

0

10

8

6

4

2

0

12

10

8

4

6

2

0

16

18

Root Tip

Root Collar

Treatment Treatment

(a)

nk nk

nk nk

nK nK

nK nK

Nk Nk

Nk Nk

NK NK

NK NK

Fig. 1. Root tip and root collar soluble sugar (a), starch (b), phenolics (c), and protein precipitable tannin (d) concentrations collectedafter three growing seasons from Douglas-fir seedlings. Treatment abbreviations are given in Table 1.

0

2

4

6

8

10

12

14

34

3

2

1

0

5

4

3

2

0

1

2

1

0

Treatment

Su

ga

r(%

)

Treatment

Sta

rch

(%)

Treatment

Ta

nn

ins

(%)

Treatment

(a) (b)

(c) (d)

nk nk

nk nk

nK nK

nK nK

Nk Nk

Nk Nk

NK NK

NK NK

Ph

en

oli

cs

(%)

Fig. 2. Root tip soluble sugar (a), starch (b), phenolics (c), and protein precipitable tannin (d) concentrations collected after threegrowing seasons from Douglas-fir seedlings.

root collar were 0.30 and 0.12 and at the root tips, 0.23 and0.61, respectively.

Foliar N levels for the low-N treatments were well belowthe “adequate” threshold, whereas high-N treatments wereabove the adequate threshold for Douglas-fir, as describedby van den Driessche (1979) and Webster and Dobkowski(1983). Plants receiving the Nk treatment had inadequate fo-liar K concentrations for growth (van den Driessche 1979;Webster and Dobkowski 1983). In addition, foliar color indi-cated K deficiency in the needles. Seedlings that receivedthe two high-N treatments had foliar K/N ratios substantiallybelow the 0.50 inadequate level described by Ingestad(1967) and Ingestad and Lund (1979). These foliar resultsare similar to those of van den Driessche and Ponsford(1995) and to the N and K foliar concentrations infield-grown Douglas-fir trees, where insufficient K was as-sociated with optimal N levels (Mika and Moore 1991).Thus, the treatment regimes used in this study successfullycreated a range of nutrient status from adequate to inade-quate, necessary for observing differences in root chemistry.

Treatment effects on root chemistry were similar for bothroot sampling locations; however, variation not directly re-lated to treatment was higher at the root collar. Different Nlevels in the treatments significantly affected seedling sizeas well as sugar and starch concentrations at both samplinglocations, and phenolic and tannin concentrations at the rootcollar. However, seedling biomass showed no correlationwith phenolic and tannin concentrations at the root tips.Seedling size differences possibly affected the mix of tissues(i.e., cambium, phloem, phellum, phelloderm, and cortex) inthe root collar bark samples and thus contributed to the rootchemistry results, while tissues sampled at the root tips wereuncorrelated with seedling size and thus more uniform incomposition. We feel that N effects on root collar phenolicand tannin concentrations derive from N effects on seedlingsize. Therefore, from the standpoint of experimental meth-odology, we believe sampling at the root tips is preferred for

experiments similar to ours, since treatment effects werepronounced and ontogenetic effects lessened. However,treatment induced changes in root collar chemistry may beas important as root tips for potential interpretation of ourresults in a field situation.

The low concentration of sugar and starch at the root tipsmay explain the lack of correlation with tannin or phenolicconcentrations. Positive correlations were shown betweenroot collar sugar concentration and tannin or phenolic con-centrations. These results are similar to those of Gebauer etal. (1998) and support the CNB hypothesis that secondarycarbon compound synthesis depends directly on nonstruc-tural carbohydrate levels. However, root collar starch con-centration was significantly negatively correlated with bothtannin and phenolic concentrations. This inverse relationshipis similar to that of Gebauer et al. (1998) for their lateralroot sampling location; however, they showed a positive re-lationship between starch and tannin for their tap root sam-ple. Perhaps the inverse relationships between starch andtannin or phenolic concentrations in our study resulted fromthe K treatment effect on tannin and phenolic concentrations.

Presumably the N and K treatments caused Douglas-firseedlings to alter carbohydrate production and subsequentphotosynthate allocation among biosynthetic pathways andseedling parts. Generally, starch was reduced in seedlings re-ceiving the high-N treatments, whereas phenolics and tan-nins were reduced in plants receiving low-K treatments. Theseedlings grown with high-N regimes had significantlylower levels of starch in both the root collar and root tipsthan in the low-N regimes. Both Wargo (1984) and Entry etal. (1991a) reported similar changes among root carbohy-drates associated with N levels in conifers. In our study, phe-nolic and tannin levels in the root collar area wereunaffected by high-N treatments. However, other studiesfound decreased phenolic levels in leaf or needle tissue withincreased N (Bryant et al. 1987; Joseph et al. 1993) or an in-crease in phenolic leaf concentration with decreased N(Poorter et al. 1997).

A K deficit for Douglas-fir in our study produced lowerconcentrations of phenolics and tannins in the roots. Potas-sium deficiencies may have affected K-controlled enzymaticactivities that affect carbon allocation to the shikimic acidpathway, which produces defensive carbon-based com-pounds (Mooney 1972). Furthermore, seedlings receiving

© 1998 NRC Canada

Shaw et al. 1571

0

0.2

0.4

0.6

0.8

1

1.2

Ph

en

olic

/Su

ga

rR

atio

Treatmentnk nK Nk NK

Fig. 3. Root tip phenolic to sugar concentration ratios collectedafter three growing seasons from Douglas-fir seedlings.Treatment abbreviations are given in Table 1.

Rootcollar

Roottips

Sugar Phenolics 0.32** 0.00nsa

Tannins 0.32** 0.01nsStarch Phenolics –0.32** 0.01ns

Tannins –0.16* 0.02nsBiomass (g/seedling) Phenolics 0.55** –0.01ns

Tannins 0.62** –0.03nsans, not significant (p > 0.10).*p < 0.05.** p < 0.01.

Table 6. Coefficient of determination (r2) between totalphenolics or protein precipitable tannin concentrations andsoluble sugars, starch, and total seedling biomass at differentroot sampling locations.

the high-N treatments were growing rapidly and may haveallocated more carbon to sugar and cellulose production andless to secondary metabolites, such as phenolics and tannins(Entry et al. 1991b). Penuelas and Estiarte (1998) speculatedabout the sources of variation that might explain differentresults among experiments dealing with elevated CO2 effectson secondary carbon-based compounds. They suggested thatthe effect of soil nutrient status could explain part of the di-versity of results in the studies they reviewed. Our resultssuggest that K status should be particularly considered asone source of variation in studies dealing with phenolics andtannins.

Wargo (1980), working with sugar maple, and Entry et al.(1991a, 1991b), working with Douglas-fir, reported that in-creased levels of glucose enable theArmillaria fungus togrow more rapidly in tree roots, making the fungus betterable to break down phenolic compounds. The effects of ex-cessive N on nutritional balance and its negative effect onplant resistance to disease have been well established.(Matson and Waring 1984; Entry et al. 1986; Ylimartimo1991). In addition, Entry et al. (1991a, 1991b) found that thephenol/sugar ratio was related to susceptibility toArmillariainfection, with low ratios (i.e., low phenolics and high sug-ars) promoting spread of infection. In this study, high N pluslow K or low K alone resulted in the lowest phenol to sugarratios. Our results suggest that the effect of low K coupledwith high N fertilization rates in field experiments (Mikaand Moore 1991), which produced an imbalance in foliarK/N ratios and subsequent high tree mortality due to rootdisease, may be explained by fertilization induced changesin root biochemistry that favored the spread of root disease.Gebauer et al. (1998) noted that the large investment ofloblolly pine (Pinus taedaL.) into root phenolics to reach ahigher phenolic/sugar ratio, and thus better resistance topathogens, suggests that belowground pathogens are impor-tant for growth and survival of loblolly pine. The responseof these same root compounds to nutrient treatments in ourstudy may indicate that defense against belowground patho-gens is equally important for successful growth of Douglas-fir. Additional field experiments testing the relationships be-tween N and K nutrition, root biochemistry, and the spreadof root pathogens are currently underway.

The effects of high and low N and K on the production ofstarch, phenolic, and protein-precipitable tannin concentra-tions have been demonstrated in this study. Imbalances be-tween N and K led to nutritional stress and secondaryproduct imbalances, which may decrease resistance to dis-ease. The foliar N and K levels in this study were similar toN and K concentrations in field-grown plants (Mika andMoore 1991; Moore et al. 1994), and our study provides aplausible explanation for the increased incidence of root dis-ease they observed after fertilizing low-K sites with high Nrates. Therefore, these relationships between N and K nutri-tion and root chemistry should provide a better understand-ing of the relationships between mineral nutrition and treeresistance to disease in a forest environment.

The support of the Stillinger Foundation and the Inter-mountain Forest Tree Nutrition Cooperative located at theUniversity of Idaho is acknowledged and greatly appreciated.

Bazzaz. F.A., Chiariello, N.R., Coley, P.D., and Pitelka, L.F. 1987.Allocating resources to reproduction and defense. BioScience,37: 58–67.

Bremner, J.M., and Mulvaney. C.S. 1982. Methods of soil analysis.Part 2. Chemical and microbiological properties. Agronomy,No. 9. pp. 595–624.

Bryant, J.P., Chapin, F.S., Reichardt, P.B., and Clausen, J.P. 1983.Response of winter chemical defense in Alaska paper birch andgreen alder to manipulation of plant carbon/nutrient balance.Oecologia,72: 510–514.

Bryant, J.P., Clausen, T.P., Reichardt, P.B., McCarthy, M.C., andWerner, R.A. 1987. Effect of nitrogen fertilization upon the sec-ondary chemistry and nutritional value of quaking aspen(Populus tremuloidesMichx.) leaves for the large aspen tortix(Choristoneura conflictanaWalker). Oecologia,73: 513–517.

Dudt, J.F., and Shure, D.J. 1994. The influence of light and nutri-ents on foliar phenolics and insect herbivory. Ecology,75:86–98.

Entry, J.A., Martin, N.E., Cromack, K., Jr., and Stafford, S. 1986.Light and nutrient limitation inPinus monticola: seedling sus-ceptibility to Armillaria infection. For. Ecol. Manage.17:189–198.

Entry, J.A., Cromack, K., Jr., Hansen, E., and Waring, R.H. 1991a.Response of western coniferous seedlings to infection byArmillaria ostoyaeunder limited light and nitrogen. Phytopa-thology, 87: 89–94.

Entry, J.A., Cromack, K., Jr., Kelsey, R.G., and Martin, N.E.1991b. Response of Douglas-fir to infection byArmillariaostoyae after thinning or thinning plus fertilization. Phyto-pathology,81: 682–689.

Garraway, M.O. 1975. Stimulation ofArmillaria melleagrowth byplant hormones in relation to the concentration and type of car-bohydrate. Eur. J. For. Pathol.5: 35–43.

Gebauer, R.L.E., Strain, B.R., and Reynolds, J.F. 1998. The effectof elevated CO2 and N availability on tissue concentrations andwhole plant pools of carbon based secondary compounds inlobolly pine (Pinus taeda). Oecologia,113: 29–36.

Hagerman, A.E. 1987. Radical diffusion method for determiningtannin in plant extract. J. Chem. Ecol.13: 437–449.

Hansen, J., and Moller, I. 1975. Percolation of starch and solublecarbohydrates from plant tissue for quantitative determinationwith anthrone. Anal. Biochem.68: 87–94.

Heath, L.S., and Chappell, H.N. 1989. Growth response to fertil-ization in young Douglas-fir stands. West. J. Appl. For.4(4):116–119.

Herms, D.A., and Mattson, W.J. 1992. The dilemma of plants: togrow or defend. Q. Rev. Biol.67: 283–335.

Huber, D.M., and Arny, D.C. 1985. Interactions of potassium withplant disease.In Potassium in agriculture.Edited by R.D.Munson. American Society of Agronomy, Madison, Wis.pp. 467–488.

Ingestad, T. 1967. Methods for uniform optimum fertilization offorest tree plants.In Proceedings, 14th IUFRO Congress, 4–9Sept. 1967, Munich, Germany. Vol. 3. pp. 265–269.

© 1998 NRC Canada

1572 Can. J. For. Res. Vol. 28, 1998

© 1998 NRC Canada

Shaw et al. 1573

Ingestad, T., and Lund, A.B. 1979. Nitrogen stress in birch seed-lings. 1. Growth technique and growth. Physiol. Plant.45:137–148.

Joseph, G., Kelsey, R.G., Moldeke, A.F., Miller, J.C., Berry, R.E.,and Wernz, J.G. 1993. Effects of nitrogen and Douglas-firallelochemicals on develpoment of the gypsy moth,Lymantriadispar. J. Chem. Ecol.19: 1245–1263.

Julkunen-Tito, R. 1985. Phenolic constituents of northern willows:methods for the analysis of certain phenolics. J. Agric. FoodChem.33: 213–217.

Kelsey, R.G., and Harmon. M.E. 1989. Distribution and variationof extractable total phenols and tannins in the logs of four coni-fers after one year on the ground. Can. J. For. Res.19: 1030–1036.

Larsson, S., Wiren, A., Lundgren, L., and Ericson, T. 1986. Effectsof light and nutrient stress on leaf phenolic chemistry inSalixdasyclados and susceptibility to Galerucella limeola(Coleoptera). Oikos,47: 205–210.

Loomis, W.E. 1932. Growth differentiation balance versus carbo-hydrate – nitrogen ratio. Proc. Am. Soc. Hortic. Sci.29:240–245.

Matson, P., and Waring. R.H. 1984. Effects of nutrients and lightlimitations on mountain hemlock: susceptibility to laminatedroot rot. Ecology,65: 1517–1524.

Marshall, J.D. 1986. Drought and shade interact to cause fine-rootmortality in Douglas-fir seedlings. Plant Soil,91: 51–60.

Mika, P.G., and Moore. J.A. 1991. Foliar potassium status explainsDouglas-fir response to nitrogen fertilization in the InlandNorthwest, USA. Water Air Soil Pollut.54: 477–491.

Mooney, H.A. 1972. The carbon balance of plants. Annu. Rev.Ecol. Syst.3: 315–346.

Moore, J.A., Mika, P.G., and Vander Ploeg, J.L. 1991. Nitrogenfertilizer response of Rocky Mountain Douglas-fir by geo-graphic area across the Inland Northwest. West. J. Appl. For.6(4): 94–98.

Moore, J.A., Mika, P.G., Schwandt, J.W., and Shaw, T.M. 1994.Nutrition and forest health.In Symposium Proceedings: InteriorCedar – Hemlock – White Pine Forest: Ecology and Manage-ment, 2–4 Mar. 1993, Spokane, Wash.Compiled by D.A.Baumgartner. Washington State University Cooperative Exten-sion Service, Pullman. pp. 173–176.

Mwangi, L.M., Lin, D., and Hubbes, M. 1990. Chemical factors inPinus strobusinhibitory to Armillaria ostoyae. Eur. J. For.Pathol.20: 8–14.

Ostrofsky, A., and Shigo, A.L. 1984. Relationship between cankersize and wood starch in American chestnut. Eur. J. For. Pathol.14: 65–68.

Parker, J., and Houston, P.R. 1971. Effects of repeated defoliationon root and root collar extractives of sugar maple trees. For. Sci.17: 91–95.

Penuelas, J., and Estiarte, M. 1998. Can elevated CO2 affect sec-ondary metabolism and ecosystem function? TREE,13: 20–24.

Poorter, H., Van Berkel, Y., Baxter, R., Den Hertog, J., Dijkstra, P.,Gifford, R.M., Griffin, K.L., Roumet, C., Roy, J., and Wong,S.C. 1997. The effect of elevated CO2 on the chemical composi-tion and construction costs of leaves of 27 C3 species. Plant CellEnviron. 20: 472–482.

Rose, R., Rose, C.L., Omi, S.K., Forry, K.R., Dural, D.M., andBigg, W.L. 1991. Starch determination by perchloric acid vs. en-zymes: evaluating the accuracy and precision of six colorimetricmethods. J. Agric. Food Chem.39: 2–11.

SAS Institute Inc. 1985. SAS user’s guide to statistics. SAS Insti-tute Inc., Cary, N.C. pp. 433–506.

Shafii, B., Moore, J.A., and Olson, J.R. 1989. Effects of nitrogenfertilization on growth of grand fir and Douglas-fir stands innorthern Idaho. West. J. Appl. For.4(2): 54–57.

Shafii, B., Moore, J.A., and Newberry, J.D. 1990. Individual treediameter growth models for quantifying within-stand responseto nitrogen fertilization. Can. J. For. Res.20: 1149–1155.

Thermo Jarrell Ash Corp. 1988. ICAP 61 operator’s manual. Janu-ary edition. No. 125791-01. Thermo Jarrell Ash Corp., Franklin,Mass.

van den Driessche, R. 1979. Estimating potential response to fertil-ization based on tree tissue and litter analysis.In Proceedings,Forest Fertilization Conference,Edited by S.P. Gessel, R.M.Kenady, and W.A. Atkinson. Inst. For. Resour. Contrib. No. 40.University of Washington, Seattle, Wash. pp. 214–220.

van den Driessche, R., and Ponsford, D. 1995. Nitrogen inducedpotassium deficiency in white spruce (Picea glauca) andEngleman spruce (Picea englemannii) seedlings. Can. J. For.Res.25: 1445–1454.

Wargo, P.M. 1972. Defoliation-induced chemical changes in sugarmaple roots stimulate growth ofArmillaria mellea. Phytopa-thology, 62: 1278–1283.

Wargo, P.M. 1980. Interaction of ethol, glucose, phenolics and iso-lates ofArmillaria mellea. Phytopathology,70: 470. (Abstr.).

Wargo, P.M. 1984. How stress predisposes trees to attack byArmillaria mellea: a hypothesis.In Proceedings, 6th Interna-tional Conference on Root and Butt Rots of Forest Trees, 25–31Aug. 1983, Melbourne, Victoria, and Gympie, Queensland, Aus-tralia. Edited byG.A. Kile. IUFRO. pp. 115–121.

Waring, R.H., and Schlesinger, W.H. 1985. Forest ecosystems. Ac-ademic Press, Harcourt Brace, Jovanovich.

Waring, R.H., McDonald, A.J.S., Larson, S., Ericsson, T., Wiren,A., Arwidson, E., Ericson, A., and Lohammar, T. 1985. Differ-ences in chemical composition of plants grown at constant rela-tive rates with stable mineral nutrition. Oecologia,66: 157–160.

Webster, S.R., and Dobkowski, A. 1983. Concentrations of foliarnutrients for trees and the dosage and frequency of fertilizertrial. Weyerhaeuser Res. Rep. No. 1. Weyerhaeuser Corp.

Ylimartimo, A. 1991. The effect of nitrogen and potassium avail-ability on Scleroderriscanker of Scots pine seedlings. Water AirSoil Pollut. 54: 307–313.