understory species interactions in mature boreal mixedwood forests

Understory species interactions in mature borealmixedwood forests

Virginia Chavez and S. Ellen Macdonald

Abstract: We explored interactions among plant growth forms in the understory of mature boreal mixedwood forests inwestern Canada by investigating the competitive influence of erect shrubs on herbs (forbs and grasses). We established 10pairs of plots; all erect shrubs were removed in one plot of each pair (removals) and left intact in the other plot (controls).Two years later, we harvested all aboveground biomass of the herbaceous layer (herb biomass: this included graminoids,forbs, trailing shrubs, and species with a woody base but not woody stems) from the 20 plots. We tested for significantdifferences in understory species biomass and composition between control and removal plots and examined the influenceof 25 environmental factors on species composition of the herbaceous layer. Competition intensity was measured by thenatural logarithm of response ratio (ln RR) index based on herb biomass. After erect shrub removal, there was a significantincrease in herb biomass, mostly due to an increase of the most common species (e.g., Cornus canadensis Linnaeus, Lin-naea borealis Linnaeus). The values of competition intensity (ln RR) varied among herb species but were, overall, posi-tive, indicating a release from competition following shrub removal. Composition of the herbaceous layer wassignificantly different between removal and control plots and was also significantly related to seven environmental factors,which explained 40% of the variation in composition. Our study suggests that there is asymmetric competition for light be-tween erect shrub and herb species in boreal ecosystems.

Key words: removal experiment, functional diversity, understory community, species interactions, dominant species, borealforest.

Resume : Les auteurs ont etudie les interactions entre les formes de croissance des plantes en sous etage de forets borealesmixtes matures dans l’ouest du Canada, en examinant l’influence competitive des arbustes dresses et des herbacees. Ils ontetabli 10 paires de parcelles, supprimant tous les arbustes dresses dans une parcelle de chaque paire (supprimes) et en lais-sant intacte l’autre parcelle (temoins). Deux ans plus tard, ils ont recolte toute la biomasse epigee de la strate herbacee(biomasse herbacee; incluant les plantes gramineennes et non-gramineennes, les arbustes rampants et les especes muniesd’une base ligneuse mais sans tige ligneuse) des 20 parcelles. Ils ont effectue des tests de differences significatives pour labiomasse des especes de sous etage et la composition entre les parcelles temoins et de suppression, pour examiner l’in-fluence de 25 facteurs environnementaux sur la composition en especes de la strate herbacee. Ils ont mesure l’intensite dela competition par le log du Rapport de Reaction (ln RR), un index base sur la biomasse herbacee. Suite a la suppressiondes arbustes dresses, on observe une augmentation significative de la biomasse herbacee, due en grande partie a une aug-mentation des especes les plus communes (e.g. Cornus canadensis Linnaeus, Linnaea borealis Linnaeus). La valeur del’intensite de la competition (ln RR) varie entre les especes herbacees, mais demeure dans l’ensemble positive, ce qui in-dique une liberation de la competition suite a l’enlevement des arbustes. La composition de la strate herbacee s’avere si-gnificativement differente, entre les parcelles avec suppression et temoins, et montre egalement une relation significativeavec sept facteurs environnementaux, lesquels expliquent 40 % de la variation de la composition. Les resultats suggerentl’existence d’une competition asymetrique pour la lumiere entre les especes arbustives dressees et herbacees, dans les eco-systemes boreaux.

Mots-cles : experience d’elimination, diversite fonctionnelle, communaute de sous etage, interactions interspecifiques, es-peces dominantes, foret boreale.

[Traduit par la Redaction]

Introduction

Interactions among plant species play an important role inregulating composition of local communities and ecosystems(Brooker 2006) and in mediating ecosystem functioningwhen a given species or functional group is lost (Dıaz et al.2003). Plant interactions entail complex combinations of

competition and facilitation among species, but the preciseeffect of species interactions on community structure andecosystem productivity is closely tied to local availability ofresources, such as light, water, and nutrients (Tilman 2007).Much of the experimental work on plant interactions hasbeen performed in synthetically assembled communitieswith a relatively low number of species with responses

Received 11 March 2010. Accepted 17 August 2010. Published on the NRC Research Press Web site at botany.nrc.ca on 1 October 2010.

V. Chavez1 and S.E. Macdonald. Department of Renewable Resources, University of Alberta, Edmonton, AB T6G 2H1, Canada.

1Corresponding author (e-mail: [email protected]).

912

Botany 88: 912–922 (2010) doi:10.1139/B10-062 Published by NRC Research Press

monitored for a few growing seasons (Dıaz et al. 2003;Jiang et al. 2009). These studies provide limited insight intothe natural dynamics of later-sucessional plant communitieswhere realistic diversity and abundance patterns are present(Thompson et al. 2005; Jiang et al. 2009).

Understory plant communities fulfill important ecologicalroles in forested landscapes (Maguire and Forman 1983;Gilliam 2007). In the boreal forest, understory shrubs andherbs represent the largest proportion of boreal plant diver-sity and provide food and habitat for wildlife species (DeGrandpre et al. 2003; Nilsson and Wardle 2005). Also,understory communities indirectly control forest succes-sional trends by altering the microclimatic conditions in theforest floor and influencing recruitment of tree species (Hartand Chen 2006). In turn, boreal trees exert important effectson the microenvironment below the canopy and are a keydriver of understory plant community composition and di-versity (Macdonald and Fenniak 2007; Chavez and Macdon-ald 2010). In this way, trees and understory plant speciesconstantly interact throughout the different stages of borealforest succession, either by competing at the seedling stage(Lieffers et al. 1993; Constabel and Lieffers 1996) or bytrees affecting microclimate conditions for the growth ofunderstory species (Macdonald and Fenniak 2007; Chavezand Macdonald 2010). Thus, knowledge of the natural dy-namics of understory plant interactions is important for abetter understanding of forest ecosystem functioning (Gil-liam and Roberts 2003). Nevertheless, very little is knownabout interactions among understory functional groups inmature boreal mixedwood forests and in boreal forests ingeneral (but see Arii and Turkington 2002; Wardle andZackrisson 2005; Hautala et al. 2007).

Boreal herbaceous communities are constantly exposedand responding to local natural disturbances, such as tram-pling by vertebrates, single tree mortality, and windthrows,which modify the composition and distribution of herb spe-cies (De Grandpre et al. 2003; Roberts and Gilliam 2003).The degree of change in the herbaceous layer after disturb-ance can be importantly influenced by the presence and den-sity of the shrub layer, as this modifies the quantity andquality of light reaching the forest floor (De Grandpre et al.2003; Roberts 2004). Light is the primary limiting resourcefor the establishment and growth of understory vascularplants across forest ecosystems, including boreal forests(Rowe 1956; George and Bazzaz 2003; Neufeld and Young2003). In size-structured communities, such as forest under-story communities, there is asymmetric competition for lightbetween taller and shorter plants because taller individualsintercept incoming light (Schwinning and Weiner 1998; Ko-hyama and Takada 2009). The shrub layer is, therefore, ex-pected to limit the biomass production, composition, anddiversity patterns of the herbaceous layer. Assessing the re-sponse of the herbaceous layer to the presence or absence oferect shrubs is important for understanding effects of small-scale disturbance in boreal forests.

In this study, we removed erect shrubs in mature borealmixedwood stands to investigate (i) whether the absence ofa given plant growth form (erect shrubs) influences biomass,diversity, and composition of another plant growth form (theherbaceous layer); and (ii) to what extent the absence oferect shrubs affects plant interactions among species in the

herbaceous layer. Herein ‘‘herbaceous layer’’ and ‘‘herb’’ re-fer to all vascular species of the forest floor including gra-minoids, forbs, trailing shrubs (e.g., Linnaea borealis L.),and species that have a woody base but herbaceous stem(e.g., Cornus canadensis L.) (see Crane 1989; Howard1993). We carried out the study in a naturally assembledcommunity because we wanted to capture the local set ofspecies as well as the natural abundance patterns occurringin mature mixedwood boreal forest stands. We hypothesizedthat increases in light to the herbaceous layer due to erectshrub removal could lead to changes in biomass and abun-dance patterns of herbs. We did not expect significantchanges in species richness because plant interactions tendto have a higher impact on species abundance (Beatty 1984;Callaway and Walker 1997) and changes in composition ofboreal understory plant communities tend to be more influ-enced by changes in species relative abundance than in spe-cies richness (Hart and Chen 2006; Chavez and Macdonald2010).

Methods

Study site and field proceduresThe study was conducted at the Ecosystem Management

Emulating Natural Disturbance (EMEND) research site (seewww.emend.rr.ualberta.ca/index.asp), located in northwestAlberta, Canada (approximate site centre: 56846’13" N,118822’28" W) in the Clear Hills Upland Ecoregion withinthe Boreal plains Ecozone (Strong and Leggat 1992; Kish-chuk 2004). Forests in the area are codominated by trem-bling aspen (Populus tremuloides Michx.), balsam poplar(Populus basamifera L.) and white spruce (Picea glauca(Moench) Voss) with lesser amounts of black spruce (Piceamariana (Mill.) B.S.P.), balsam fir (Abies balsamea (L.)Mill.), lodgepole pine (Pinus contorta Dougl.), and paperbirch (Betula papyrifera Marsh.). Soils have developed onfine-textured glacial till or glaciolacustrine deposits and soiltypes include (generally) well-drained Luvisols (Dark GreyLuvisol, Orthic Grey Luvisol) with limited occurrences ofBrunisolic, Gleysolic, and Solonetzic soils (Kishchuk 2004).Site elevation ranges from 677 to 880 m above sea level.The region has a boreal climate with cold winters (meanJanuary temperature: –16.6 8C), warm summers (mean Julytemperature: 16 8C) and has a mean annual precipitation of402 mm, most of which is received during the summer (En-vironment Canada 2008).

We performed this study in two forest stands classified as‘‘mixed’’ at the stand (polygon) scale by forest inventories;i.e., they had between 40% and 60% of canopy cover ofconifer (mostly white spruce) and broadleaf trees (mostlytrembling aspen). The two stands were approximately 10 hain size, 100 years of age, of fire origin and previously un-managed. In June 2004, we established 20 (10 pairs) 5 m di-ameter circular plots. Plot selection was based upon thefollowing criteria: (i) all pairs were under mixedwood cano-pies having at least 40% and no more than 60% of conifertrees in the plot (based on both density and canopy cover);(ii) each pair had the same dominant shrubs — mainly Vi-burnum edule (Michx.) Raf. (low bush cranberry) and Rosaacicularis Lindl. (prickly rose) — with similar percentagecover and height (approx. 1.50 m); (iii) the plots of each

Chavez and Macdonald 913

Published by NRC Research Press

pair were approximately 2 m apart from one another; and(iv) pairs were at least 30 m apart from one another and atleast 50 m away from forest edges or cut lines. One plot ofeach pair was randomly selected to be the treatment or ‘‘re-moval’’ plot, in which all the erect shrub species wereclipped and removed from the plots in June of 2004. Theother plot was used as a ‘‘control’’, in which erect shrubswere left intact. We manipulated growth forms or functionalgroups rather than species richness because manipulation offunctional groups often provides better insight into the ef-fects that non-random species loss have on communitystructure than does manipulation of species richness (Dıazet al. 2003; Balvanera et al. 2006), although opinions aredivided on this issue (Balvanera et al. 2006).

At each plot, we made visual estimates of percent cover(1%–100%) for each vascular plant species in the understory(shrubs + herbs) within a 3-m-diameter circular subplot,which was further divided into four quarters to increase ac-curacy of cover estimates. Nomenclature followed Moss(1983). In 2004, visual cover estimates were made prior toshrub removal and in 2006, prior to the harvest of herb bio-mass. In 2006, we harvested all the aboveground parts of theherbaceous layer from the 10 pairs. We did not collect anyof the orchid species (Habenaria spp., Goodyera repens R.Br., and Calypso bulbosa (L.) Oakes) because of concernsfor their conservation. Moreover, orchids were quite infre-quent, contributing little to overall biomass. The harvestedplants were classified by species, dried with electrical fansin a research trailer while in the field, and then oven-driedat 64 8C for 24 h (in the Natural Resources Analytical Lab-oratory (NRAL) of the University of Alberta) and weighedto estimate their biomass. We measured the following envi-ronmental characteristics within the 5 m circular plots.

Canopy compositionAll canopy trees (trees with diameters at breast height

(DBH) > 5 cm) were counted and their DBH (1.3 m) meas-ured to calculate basal area (BA) for conifer (BAC) andbroadleaf (BAD) trees separately and for the two together(BAT).

Ground coverWe estimated the percentage of ground cover of (i)

downed fine woody debris (pieces <8 cm diameter), (ii)downed coarse woody debris (CWD) (pieces >8 cm diame-ter), (iii) mosses, and (iv) litter.

Edaphic factorsAt the centre of each plot, we buried nylon bags with

45 mL of Amberlite IR-150 anion–cation exchange resin(J.T. Baker) within the mineral soil layer in June of 2005.Bags were retrieved after a period of 2 months and extractedas described by Thiffault et al. (2000). We then analyzedthem using a Technicon AutoAnalyzer to obtain the concen-trations of available PO4

3–, NO3– and NH4

+, and an atomicabsorption spectrophotometer to obtain the concentrationsof Ca2+, Mg2+ and K+ (in the NRAL). In August of 2005,we measured the depth of the litter (L) and organic layers(FH) at the plot centre and collected samples of organic(FH) and mineral layers (~20 cm depth). We measured pHin a paste of a 1:2 soil:water mixture for mineral soil and

1:4 soil:water for organic using a Fisher AR20 pH meterwith glass. Lastly, we measured soil moisture and tempera-ture once over the growing season in August of 2005 at theplot centre by means of time domain reflectometry and athermocouple respectively; the measures were taken at least24 h after a major precipitation event.

Incoming lightAt the centre of each plot, we took two hemispherical

canopy photographs at two different heights during both theleaf-on (August) and leaf-off (November) periods in 2005.To fully capture the influence of shrub canopy, one photowas taken at 50 cm (below shrub layer) and the other at1 m above the ground (at the shrub layer). Photographswere taken approximately 1 h after dawn, 1 h before dusk,or on overcast days to ensure uniform sky conditions and toavoid direct insolation (Comeau et al. 1998; Gendron et al.1998). Images were analyzed using SLIM (Spot Light Inter-cept Model) version 2.2e (Comeau et al. 1998, 2003) soft-ware, which takes into account seasonal sun angle and dataon cloud cover and opacity for the specific geographic loca-tion to calculate the percentage of direct, diffuse, and total(average of diffuse and direct) solar radiation.

Data analysis

Species diversity and abundanceSpecies diversity per plot for herbs was assessed using the

Shannon’s (H’) and Simpson’s (1 – D) diversity indices, andspecies richness (S). The Shannon’s index was calculatedfrom the equation

H 0 ¼ �XS

i¼1

pi ln pi

where pi is the proportional abundance of the i th species,and the Simpson index from the equation

D ¼XS

i¼1

niðni � 1ÞNðN � 1Þ

, where ni is the abundance (cover) of the ith species perplot and N is the total abundance per plot (sum of cover va-lues). For both, S is the number of species per plot (Ma-gurran 2004). We calculated these using biomass data asmeasures of species abundance. All the wintergreens (Pyrolaspp.) were grouped together, as were violets (Viola spp.),horsetails (Equisetum spp.), and grasses (Calamagrostis ca-nadensis Michx. Nutt. and Elymus innovatus Beal).

To test for differences in diversity between control and re-moval plots, we used a randomized complete block analysisof variance (ANOVA) by means of PROC MIXED (SASversion 9.1, Cary, N.C.) with the following model: Yijk =m + Ti + Sj + (TS)ij +P(S)jk + eijk, where Yijk = diversitymeasure for the ith treatment on the jth pair at the kth stand;m = overall mean; Ti = treatment (i = 1–2; fixed); Sj = stand(block) (j = 1–2; random); (TS) ij = interaction betweenstand and treatment; P(S) jk = pair nested within stand (k =1–10); and e = experimental error. Residuals were tested toensure they met the assumptions of normality (Kolmogorov–Smirnov statistic) and homogeneity of variance (Levene’s

914 Botany Vol. 88, 2010


test) (PROC GLM; SAS version 9.1; SAS Institute Inc.2003).

We sampled a total of 42 species in the herbaceous layer(listed in Appendix A, Table A1) of which seven taxa (sixspecies and the three Pyrola spp. that were grouped duringthe harvesting of biomass) were present in all 10 pairs ofplots. These were designated as ‘‘common’’ species (listedin Table 1). Three of these (Cornus canadensis, Mertensiapaniculata (Aiton) G.Don, and Rubus pubescens Raf.) wereconsidered to also be dominant, based on their high biomass.We assessed the effect of treatment (erect shrub removal) onthe biomass of each of these seven taxa separately and onbiomass of the seven together. To test whether the abun-dance of these seven taxa was significantly different be-tween control and removal plots prior to shrub removal, wecompared pretreatment (2004) cover values for each of theseven, and for all seven combined. There were not signifi-cant differences in species cover in 2004, before implemen-tation of the removal treatment. We used the ANOVAdescribed above to perform these analyses.

Using the same ANOVA model, we tested whether therewas a difference in environmental conditions between thepaired control and removal plots. We compared mean valuesfor the 25 measured environmental characteristics: soil nu-trients (PO4

3 -, NO3–, NH4

+, Ca2+, Mg2+, and K+); edaphicfactors (litter and organic layer depth, moisture, and temper-ature); ground cover percentage (litter, moss, coarse and finewoody debris); basal area (conifer, broadleaf, and total); di-rect, diffuse, and total light during leaf-on and leaf-off peri-ods; and total annual light.

Plant competition intensityWe used the natural logarithm of response ratio (ln RR)

competition intensity index to measure competitive interac-tion between shrub and herb species using the herb biomassdata (collected in 2006, 2 years after shrub removal) in con-trol and removal plots. The index was calculated using thefollowing equation: ln[(X removal)/(X control)], where X repre-sents the aboveground herb biomass in the removal and con-trol plots (Goldberg et al. 1999). We selected this indexbecause it is symmetrical for competitive and facilitative in-teractions and does not set a ceiling value on the maximumpossible competition intensity (Goldberg et al. 1999). Posi-tive ln RR values indicated that there was greater herb bio-

mass after the shrub removal treatment, suggesting acompetitive influence of erect shrubs on herbs. Conversely,negative ln RR values occurred when there was lower herbbiomass after treatment, indicating a facilitative relation be-tween erect shrubs and herbs (Goldberg et al. 1999). Formathematical properties of ln RR, see Goldberg et al.(1999) and Oksanen et al. (2006).

For each pair (control–removal) of plots, we calculatedthe following ln RR indices: (i) an overall index includingall herb species, (ii) an index including the seven most com-mon herbs, and (iii) a separate index for each of the sevencommon herbs. Then, to assess the potential influence ofthe removal of erect shrubs on the competition intensityamong herb species, we regressed the ln RR of each of theseven common herbs as a function of the following varia-bles: (i) total herb biomass (all herb species combined), (ii)total biomass of each of the seven common herbs (excludingthe biomass of the ‘‘response’’ taxa), and (iii) biomass ofeach of the seven common taxa regressed on each of theothers separately (e.g., the InRR value of species Y in a re-moval plot as a function of the biomass of species X in thatplot). All the regressions were performed in SAS version 9.1(SAS Institute Inc. 2003) using PROC REG.

Species compositionTo assess the influence of the removal treatment on herb

species composition, we compared the removal and controlplots by means of a permutation-based nonparametric MAN-OVA (PerMANOVA; Anderson 2001). We used a random-ized complete block design with pairs as blocks and Bray–Curtis as a distance measure. To identify herbs that couldbe used as indicators of either removal or control plots, weperformed an indicator species analysis (Dufrene and Legen-dre 1997; McCune and Grace 2002). These analyses wereperformed using PC-Ord for Windows (version 4.5)(McCune and Mefford 1999). To assess whether the inten-sity of competition from which the community was releasedby the shrub removal treatment affected the extent of changein community composition, we calculated Bray–Curtis eco-logical distances between the paired control and removalplots with the biomass data for the 42 species using thefunction vegdist in the vegan package of R (version 2.3.1.)(Oksanen et al. 2009). Then, we regressed the Bray–Curtisvalues versus the InRR values of the 42 species together, as

Table 1. Results of mean biomass (g of dry biomass) for the control and removal treatments of the dominant (D) andcommon (C) species at the time of harvest of herb biomass in 2006.

Species Control Removal p value In RRa

Cornus canadensis (D&C) 43.269 (±13.93)b 76.642 (±33.66) 0.03 0.53 (±0.40)b

Epilobium angustifolium (C) 19.902 (±14.22) 23.677 (±17.15) N.S. 0.14 (±0.88)Linnaea borealis (C) 9.963 (±4.56) 29.524 (±12.11) 0.02 1.24 (±1.30)Mertensia paniculata (D&C) 30.283 (±11.72) 17.622 (±17.62) N.S. –0.15 (±0.42)Mitella nuda (C) 9.332 (±4.47) 12.706 (±5.09) N.S. 0.08 (±0.31)Pyrola spp. combined (C) 6.261 (±3.37) 8.868 (±4.17) N.S. 0.12 (±0.99)Rubus pubescens (D&C) 34.507 (±15.38) 41.123 (±17.76) N.S. 0.17 (±0.52)All common & dominant species 151.93 (±36.26) 221.93 (±40.77) 0.008 0.38 (±0.24)All 42 species 237.747 (±13.93) 310.891 (±69.65) 0.032 0.30 (±0.29)

Note: Herbs listed were found in all of the pairs of sampling plots.aln RR values representing competition intensity between shrub and herb species.bThe value within brackets is the ±95% confidence limit.



well as the seven common herbs (PROC REC SAS version9.1; SAS).

To relate environmental characteristics to understory spe-cies composition, we used a constrained ordination. Wechose a distance-based redundancy analysis (dbRDA)(Legendre and Anderson 1999; McArdle and Anderson2001) because it can be used with non-Euclidean distances.Also, significance testing is done using non-parametric per-mutations and thus there is no assumption of normality. ThedbRDA was performed by first calculating a Bray–Curtis(Sørensen) ecological distance on the (untransformed) herbbiomass per plot to data. A principal coordinate analysiswas then used to extract the principal coordinates of the dis-tance matrix, correcting for negative eigenvalues. We usedthis resulting matrix as the species data in a redundancy anal-ysis (RDA), wherein the environmental matrix contained con-tinuous environmental variables and orthogonal dummyvariables coding for the two treatments (removal and control).

A total of 25 environmental variables (mentioned above)were tested for inclusion as an explanatory variable in thisanalysis. Because the number of environmental and foreststructure variables exceeded the number of plots, it was notpossible to test all the variables together. Thus, we madesubsets of environmental variables and conducted separatedbRDAs on these subsets, using stepwise-forward selectionwith Monte Carlo permutations to test each variable for sig-nificance. For our final analysis, we included only the varia-bles that were constantly significant (p < 0.05). To considerthe paired design of the study, we used each pair as a block(coded as a co-variable). Finally, for interpretation purposes,we used the original species data (not the distance-based)matrix as a supplemental dataset. Also, to investigate the ef-fect of shrub removal on community composition, we ran adbRDA including a dummy variable coding for treatment as

a single environmental variable. We performed these analy-ses in CANOCO (version 4.5) (ter Braak and Smilauer2002).

Results

Species diversity & abundanceNone of the assessed diversity measures (1–D, 1/D, E1/D,

H’, and S) were significantly different between the controland removal plots before or after the imposed treatment (re-sults not shown).

Total herb biomass and the biomass of the seven commonherbs combined was significantly higher (p < 0.032 and p <0.008, respectively) in removal plots than in control plots(Table 1). The seven common taxa in the herbaceous layertogether had higher biomass than all the rest of the herba-ceous layer species combined for both the removal (p <0.0001) and control (p < 0.028) plots. Collectively, theseseven taxa accounted for more than 60% of total herbaceouslayer biomass in control plots and over 70% in removalplots (Table 1). Mean biomass of Linnaea borealis was sig-nificantly higher in the removal plots than in control plots;this was also the case for Cornus canadensis. The abun-dance of the other five common taxa was not significantlydifferent between treatments.

Diffuse, direct, and total incoming light at 50 cm abovethe ground during the leaf-on period were significantlylower in control plots than in the removal plots, as was totalannual light at 50 cm above the ground (Table 2). Indeed,light at 50 cm aboveground in the removal plots increasedto the levels recorded in the control plots at 100 cm. Therest of the measured environmental variables (basal area,ground cover, edaphic factors) were not significantly differ-ent between control and removal plots (results not shown).

Table 2. Mean values (95% confidence interval) for incoming light measured at 50 cm and 1 m above the ground for controland clipped plots.

Light (PPDF, mmol�s–1�m2) Control 50 cm Control 1 mClipped50 cm Clipped 1 m p value

Diffuse leaf-on* 28.14 B (1.59) 33.84 A (1.34) 32.99 A (1.63) 32.69 A (1.65) < 0.0001Direct leaf-on 19.94 B (2.94) 25.38 A (2.9) 26.93 A (2.88) 26.41 A (2.58) 0.0016Total leaf-on 24.3 B (1.92) 29.91 A (1.7) 30.16 A (1.75) 29.82 A (1.65) < 0.0001Diffuse leaf-off 42.51 (2.07) 43.65 (2.11) 43.31 (2.3) 43.04 (2.57) N.S.Direct leaf-off 23.52 (4.32) 22.42 (3.55) 25.4 (3.74) 26.05 (4.27) N.S.Total leaf-off 34.98 (2.29) 35.23 (1.92) 36.2 (2.38) 36.29 (2.51) N.S.Total annual (diffuse & direct

with leaf-on & leaf-off)29.64 B (1.34) 32.57 A (1.24) 33.18 A (1.78) 33.05 A (1.74) 0.0031

*Values for a given variable that have different superscript letters were significantly different based on least squared means comparisons (ata < 0.05).

Table 3. Results of the analysis of the natural logarithm response ratio (ln RR) of individual target plants as a function of the biomass ofco-occurring species by means of linear regressions.

Target species (In RR)Co-occurring speciesbiomass R2 Adjusted R2 SS B* F value p value

Linnaea borealis Mitella nuda 0.52 0.46 6.62 –0.12 8.75 0.018Linnaea borealis Pyrola spp. 0.47 0.41 6.06 0.14 7.26 0.027Pyrola spp. Mitella nuda 0.69 0.65 9.29 –0.14 16.32 0.004

Note: Only the significant regressions are shown.*Slope of regression.

916 Botany Vol. 88, 2010


Plant competitionThe average of the ln RR value for all the herbs combined

was positive (0.30), indicating a slight release from competi-tion due to shrub removal (Table 1). The average ln RRvalue for the seven dominant species together was also pos-itive (0.38), and so were the average ln RR values for Lin-naea borealis and Cornus canadensis (1.24 and 0.53,respectively), indicating a release from competition(Table 1). The ln RR values for the rest of the herb specieswere close to zero, indicating that the removal treatment didnot have a detectable effect on them. The ln RR values ofLinnaea borealis (p = 0.02; adjusted R2 = 0.46) and of Py-rola spp. (p = 0.004; adjusted R2 = 0.65) were negatively re-lated to the biomass of Mitella nuda (Table 3). This suggeststhat when the biomass of co-occurring M. nuda was higher,L. borealis and Pyrola spp. showed less release due to shrubremoval. Conversely, the ln RR values for Linnaea borealiswere positively related to the biomass of co-occurring Py-rola spp. (p = 0.03; adjusted R2 = 0.41) (Table 3). The lattersuggests that L. borealis showed more release from competi-tion following shrub removal in plots that also had higherbiomass of Pyrola spp. The rest of the regressions betweenln RR values and biomass of individual herbs were non-significant (results not shown).

Species compositionBased on the PerMANOVA, understory community com-

position was significantly different between removal andcontrol plots (p = 0.03). The indicator species analysis iden-tified that Cornus canadensis (species indicator values =63.9) and Linnaea borealis (species indicator values = 74.8)were significantly (p < 0.02) more abundant and frequent in

removal plots. Competition intensity (ln RR) calculated withall the herbs and the seven dominant herb species combinedwere positively related to the Bray–Curtis ecological dis-tance between paired control–removal plots (p = 0.003; ad-justed R2 = 0.68 and p = 0.002; adjusted R2 = 0.67,respectively) (Table 4). This indicated that the plots whereincompetition was strongest showed the greatest difference incommunity composition after shrub removal.

The following seven environmental variables were signifi-cantly related to understory community composition (pre-sented in the order of forward selection): Total light duringthe leaf-on period at 1 m, which was positively correlated tothe first axis; treatment (control vs. removal plots) beingnegatively correlated to the second axis; soil temperatureand total basal area, which were positively correlated to thefourth axis; and total annual light at 1 m, total light duringthe leaf-on period at 50 cm, and soil Mg2+, which were pos-itively correlated to the first axis (Table 5). Together, thesevariables explained 40% of the variation in understory com-position. Treatment by itself explained only 6.3% of the var-iation in herb composition.

DiscussionWe assessed interactions among forest understory vascu-

lar plants in mature and unmanaged boreal mixedwood for-ests by investigating whether the removal of erect shrubswould influence the biomass, composition, and diversity ofthe herbaceous layer. We observed significant effects oferect shrub removal on herb biomass and community com-position, and these were mainly because of the increases inabundance of dominant and common herbs. In general, ourresults suggest that there is asymmetric competition for light

Table 4. Results of the analysis of the effect of competition intensity (ln RR) for all species and for the seven most common taxa of theherbaceous layer on the Bray–Curtis distance of species composition between the paired control–clipped plots.

Dependent variable Independent variable R2 Adjusted R2 SS B* F value p valueBray–Curtis All herbs (ln RR value) 0.70 0.65 0.06 0.19 17.60 0.003Bray–Curtis Seven dominant taxa (ln RR

value)0.71 0.67 0.06 0.24 19.66 0.002

*Slope of regression.

Table 5. Results of redundancy analysis showing the environmental variables that were significantly associated with herb speciescomposition (p < 0.05).

Variable Axis 1 Axis 2 Axis 3 Axis 4

Trace: 0.408Eigenvalues* 0.118 0.091 0.067 0.053Total light leaf-on at 1m 0.724 –0.455 0.145 –0.267Treatment –0.149 –0.605 –0.117 0.124Soil temperature 0.412 –0.179 0.537 0.554Total basal area –0.196 0.087 0.275 0.511Total annual light at 1 m 0.814 –0.229 –0.026 –0.210Total light leaf-on at 50 cm 0.475 –0.125 0.154 –0.337Mg2+ 0.337 0.153 –0.140 0.306

Note: Results were determined using stepwise-forward selection of variables in a distance-based redundancy analysis. Presented are the interset cor-relations (Pearson) of significant variables, in the order of the forward selection. The trace value (sum of all the canonical eigenvalues) and the eigen-values of the first four axes are also presented.

*Axis 1 and all combined axes are significant at p = 0.002.



between erect shrubs and herb species. As such, interactionsamong these two growth forms or functional groups likelyhave a significant effect on understory community structurein mature boreal mixedwood forests. Our study also con-firms the significant contribution that dominant and commonplant species make to plant community structure and to bio-mass recovery after a functional group is removed (Smithand Knapp 2003; Bret-Harte et al. 2008).

There are important discrepancies between plant interac-tions observed in natural communities and those observedin synthetic plant assemblages (Dıaz et al. 2003; Balvaneraet al. 2006). Plant interactions in natural communities are in-fluenced by assembly processes such as resource heteroge-neity, variation in dispersal strategies, and phenologicalstages across the present species as well as realistic abun-dance species distributions (Dıaz et al. 2003; Balvanera etal. 2006; Jiang et al. 2009). As such, the influence that envi-ronmental heterogeneity and dominance patterns exert onplant species interactions in the field is becoming a focus ofresearch, particularly in heterogeneous ecosystems, ratherthan an unwanted source of variance (Dıaz et al. 2003; Jianget al. 2009). Apart from our study, only a few others haveexamined the dynamics of understory plant interactions inboreal forests through removal experiments in natural com-munities (e.g., Arii and Turkington 2002; Wardle and Zack-risson 2005; Hautala et al. 2007). Through these removalstudies, understanding has increased regarding the rapid re-sponse of dominant species to the removal of co-occurringspecies, as well as the important role that dominant andcommon species play in boreal ecosystem recovery after anentire functional group or growth form is removed (Wardleand Zackrisson 2005; Hautala et al. 2007).

Our study also points to the importance of common anddominant species for biomass recovery in boreal forests.The significant difference in herbaceous species compositionbetween removal and control plots was largely attributableto rapid changes in abundance of the most common taxa,particularly of Linnaea borealis and Cornus canadensis, thetwo species that were significant indicators of removal plots.This further confirms the importance of considering abun-dance patterns, not just richness, for the assessment of bor-eal understory communities (Chavez and Macdonald 2010).Responses of less common or subordinate herb species toshrub removal may have been limited by their lower abilityto acclimatize to the changes in micro-environmental condi-tions (i.e., increases in soil temperatures and evapo-transpiration rates), by their morphological and phenologicaldifferences, or by competitive suppression by the dominantherb species (Symstad and Tilman 2001). These findingssupport the idea that dominant species tend to confer short-term resistance to reduction in ecosystem functions whenother community members are lost (Smith and Knapp2003). These results are further supported by other removalstudies carried out in ecosystems where seed recruitment isless important for community assemblage, such as the tundra(Bret-Harte et al. 2008) and seasonally dry forests(D’Antonio et al. 1998); these studies have also reportedthat empty spaces left by the removal of co-occurring spe-cies tend to be occupied by common species that are alreadydominant and that expand though vegetative growth (Dıaz etal. 2003).

Flowering frequency is low and suitable microsites forseed germination are scarce in boreal forests; thus, seed re-cruitment is not as important as vegetative growth for siteoccupancy in the understory (Økland 1995). Clonal growthimportantly influences interactions among plants (Oborny etal. 2000; Oborny and Kun 2002; Kun and Oborny 2003).Through physiological integration, interconnected rametsshare their surplus resources (i.e., water, soil nutrients, andphotoassimilates), increasing the probability of survival andreproduction among ramets of the same genet (Oborny etal. 2000; Oborny and Kun 2002; Kun and Oborny 2003). Ahigher degree of ramet integration tends to confer highercompetitive advantage to herb species inhabiting spatiallyheterogeneous systems, where resources are patchily distrib-uted (Stuefer et al. 1996; Oborny et al. 2000) as they typi-cally are in boreal forests (Carleton and Maycock 1980;Chavez and Macdonald 2010). In this study, the significantbiomass increase of the stoloniferous forb Linnaea borealisafter shrub removal was likely influenced by its ability toshare resources among its clonal fragments. In patches withhigher light availability in the forest floor, the ramets ofL. borealis increase their branching frequency to enhancecarbon assimilation, provided that light availability remainsfavourable for 2–3 years (Niva et al. 2006), which was theelapsed time after treatment in our study.

Light is an important limiting resource for the establish-ment and growth of boreal understory vascular plants(Rowe 1956; De Grandpre et al. 2003). Although competi-tion for light is expected to have considerable influence onboreal understory communities, there is some controversyregarding the importance of light competition on and amongclonal plants (Pennings and Callaway 2000; Choler et al.2001). Plant competition for light is generally thought to beasymmetric (Schwinning and Weiner 1998). However, thesymmetry or asymmetry of light competition among clonalplants may depend on height differences among ramets andthe degree of resource translocation among them (Schwin-ning and Weiner 1998; but see de Kroon et al. 1992). Thesignificant increase in herb biomass observed in this studyafter shrub removal suggests that there is an asymmetriccompetition for light between boreal erect shrubs and herbs.We did not manipulate the herb layer to assess competitionamong boreal herbs. Nonetheless, we found some evidencefor a competitive effect of Mitella nuda on L. borealis andPyrola spp. in that the latter two showed less release fromcompetition following shrub removal when biomass of Mi-tella nuda was high. These clonal, small-leafed species seg-regate in the lowest vertical strata of the forest, expandingmore horizontally than vertically. They thus exploit resour-ces from similar niche space and may compete morestrongly among themselves than with taller forb species.

The results of this study suggest that species interactionsplay an important role in influencing plant community struc-ture, in terms of herb abundance and composition, in borealmixedwood forests. However, stand density, incoming light,soil moisture, and nutrients all together explained a largerproportion of the variation in herbaceous composition thandid the shrub removal treatment; this emphasizes the stronginfluence that environmental heterogeneity exerts on borealplant communities (Carleton and Maycock 1980; Macdonaldand Fenniak 2007). The significant effect of light at the

918 Botany Vol. 88, 2010


shrub level (1 m from the forest floor) and below it (50 cmfrom the forest floor) highlights the importance that bothcanopy trees and shrubs have in mediating the amount of in-coming light that reaches the herbaceous layer (20 cm fromthe forest floor) (Constabel and Lieffers 1996; Hart andChen 2006) and the effect of this on understory plant com-munity composition.

Competition for water and nutrients can potentially limitthe growth of forest herbs (Hicks and Turkington 2000; An-derson 2003; Neufeld and Young 2003). Because shrubs im-mobilize large quantities of nutrients in the summer, theremoval of erect shrubs could have increased nutrient avail-ability for herbs (Anderson 2003). We did not assess below-ground interactions and did not detect a significantdifference in moisture or nutrients after shrub removal.

Herb response to erect shrub removal was speciesspecific, and increases in herb cover occurred mostly byvegetative growth of existing individuals. This response co-incides with that seen after microscale disturbance events inwhich the soil layer is not severely affected (e.g., tramplingand browsing by vertebrates; small wind storms) (Roberts2004). Changes in plant communities due to neighbour re-moval are thought to be slow in boreal forests (Bryant et al.1983). The significant changes in herb biomass and speciescomposition that we detected only 2 years after shrub re-moval, however, support the idea that responses to the re-moval of a growth form or functional group can be rapid innorthern forests (Leniere and Houle 2009), at least for dom-inant and common species. Also, it supports the idea thatunderstory herbs have the capacity to quickly respond tochanges in their environment following disturbance (Leniereand Houle 2009). In comparison to other forest ecosystems,boreal forests have harsher climatic conditions and lower re-source availability (Hicks and Turkington 2000; Arii andTurkington 2002). Thus, the results of this study do not sup-port the traditional idea that herbaceous communities in theforest floor of harsh environments are stress tolerators andlack the ability to rapidly or positively react to increases inresource availability (Grime 1977, 2001). Our results agreewith Arii and Turkington (2002) in that in boreal forests,not all understory herbs can be classified as stress-tolerators.

In accordance with theory regarding the influence of hab-itat heterogeneity and patch dynamics on clonal plants (Kot-liar and Wiens 1990; Oborny et al. 2000) and results fromthis and other field studies (Niva et al. 2006; Chavez andMacdonald 2010), we suggest that the matrix of contrastinglight patches is important for mediating plant–plant interac-tions in boreal mixedwood forests. Overall, this study indi-cates that interactions among understory plant species playan important role in structuring boreal understory commun-ities. Changes in the natural dynamics of these interactionsmay indirectly modify the natural structure of boreal under-story communities.

AcknowledgementsThis study was supported by grants from the Alberta Con-

servation Association (through the Challenge Grants in Bio-diversity, University of Alberta) and the Natural Sciencesand Engineering Research Council (Canada). V.C. was sup-ported partly by CONACYT: Consejo Nacional de Ciencia yTecnologıa, Mexico (the Mexican National Council of Sci-

ence and Technology). We thank Danielle Koleyak, NicoleMathew, Peter Presant, Gina Sage, and Penny Wizniuk fortheir valuable assistance in the field.

ReferencesAnderson, M.J. 2001. A new method for non-parametric multivari-

ate analysis of variance. Austral Ecol. 26(1): 32–46. doi:10.1046/j.1442-9993.2001.01070.x.

Anderson, W.B. 2003. Interactions of nutrient effects with otherbiotic factors in the herbaceous layer. In The herbaceous layerin forests of eastern North America. Edited by F.S. Gilliam andM.R. Roberts. Oxford University Press, New York. pp. 91–101.

Arii, K., and Turkington, R. 2002. Do nutrient availability andcompetition limit plant growth of herbaceous species in the bor-eal forest understory? Arct. Antarct. Alp. Res. 34(3): 251–261.doi:10.2307/1552482.

Balvanera, P., Pfisterer, A.B., Buchmann, N., He, J.S., Nakashi-zuka, T., Raffaelli, D., and Schmid, B. 2006. Quantifying theevidence for biodiversity effects on ecosystem functioning andservices. Ecol. Lett. 9(10): 1146–1156. doi:10.1111/j.1461-0248.2006.00963.x. PMID:16972878.

Beatty, S.W. 1984. Influence of microtopography and canopy spe-cies on spatial patterns of forest understory plants. Ecology,65(5): 1406–1419. doi:10.2307/1939121.

Bret-Harte, M.S., Mack, M.C., Goldsmith, G.R., Sloan, D.B., De-marco, J., Shaver, G.R., Ray, P.M., Biesinger, Z., and Chapin,F.S., III. 2008. Plant functional types do not predict biomass re-sponses to removal and fertilization in Alaskan tussock tundra.J. Ecol. 96(4): 713–726. doi:10.1111/j.1365-2745.2008.01378.x.PMID:18784797.

Brooker, R.W. 2006. Plant-plant interactions and environmentalchange. New Phytol. 171(2): 271–284. doi:10.1111/j.1469-8137.2006.01752.x. PMID:16866935.

Bryant, J.P., Chapin, F.S., III, and Klein, D.R. 1983. Carbon/nutri-ent balance of boreal plants in relation to vertebrate herbivory.Oikos, 40(3): 357–368. doi:10.2307/3544308.

Callaway, R.M., and Walker, L.R. 1997. Competition and facilita-tion: a synthetic approach to interactions in plant communities.Ecology, 78(7): 1958–1965. doi:10.1890/0012-9658(1997)078[1958:CAFASA]2.0.CO;2.

Carleton, T.J., and Maycock, P.F. 1980. Vegetation of the borealforests south of James Bay: non-centered component analysis ofthe vascular flora. Ecology, 61(5): 1199–1212. doi:10.2307/1936838.

Chavez, V., and Macdonald, S.E. 2010. The influence of canopypatch mosaics on understory plant community composition inboreal mixedwood forests. For. Ecol. Manage. 259(6): 1067–1075. doi:10.1016/j.foreco.2009.12.013.

Choler, P., Michalet, R., and Callaway, R.M. 2001. Facilitation andcompetition gradients in alpine plant communities. Ecology,82(12): 3295–3308. doi:10.1890/0012-9658(2001)082[3295:FACOGI]2.0.CO;2.

Comeau, P.G., Gendron, F., and Letchford, T. 1998. A comparisonof several methods for estimating light under a paper birch mix-edwood stand. Can. J. For. Res. 28(12): 1843–1850. doi:10.1139/cjfr-28-12-1843.

Comeau, P.G., Macdonald, R., and Bryce, R. 2003. SLIM (SpotLight Intercept Model). Version 2.2d. B.C. Ministry of Forests,Victoria, B.C.

Constabel, A.J., and Lieffers, V.J. 1996. Seasonal patterns of lighttransmission through boreal mixedwood canopies. Can. J. For.Res. 26(6): 1008–1014. doi:10.1139/x26-111.

Crane, M.F. 1989. Cornus canadensis. In Fire Effects Information



System [online]. U.S. Department of Agriculture, Forest Service,Rocky Mountain Research Station, Fire Sciences Laboratory(Producer). Available from www.fs.fed.us/database/feis/ [Ac-cessed February 2010].

D’Antonio, C.M., Hughes, R.F., Mack, M., Hitchcock, D., and Vi-tousek, P.M. 1998. The response of native species to removal ofinvasive exotic grasses in a seasonally dry Hawaiian woodland.J. Veg. Sci. 9(5): 699–712. doi:10.2307/3237288.

De Grandpre, L., Bergeron, Y., Nguyen, T., Boudreault, C., andGrondin, P. 2003. Composition and dynamics of the understoryvegetation in the boreal forest of Quebec. In The herbaceouslayer in forests of eastern North America. Edited byF.S. Gilliam and M.R. Roberts. Oxford University Press, NewYork. pp. 238–261.

de Kroon, H., Hara, T., Kwant, R., and Kwant, R. 1992. Size hier-archies of shoots and clones in clonal herb monocultures: doclonal and non-clonal plants compete differently? Oikos, 63(3):410–419. doi:10.2307/3544967.

Dıaz, S., Symstad, A.J., Chapin, F.S., III, Wardle, D.A., and Huen-neke, L.F. 2003. Functional diversity revealed by removal ex-periments. Trends Ecol. Evol. 18(3): 140–146. doi:10.1016/S0169-5347(03)00007-7.

Dufrene, M., and Legendre, P. 1997. Species assemblages and indi-cator species: the need for a flexible asymmetrical approach.Ecol. Monogr. 67(3): 345–366. doi:10.1890/0012-9615(1997)067[0345:SAAIST]2.0.CO;2.

Environment Canada. 2008. Canadian climate normals 1971–2000:Peace River, Alberta [online]. Available from www.climate.weatheroffice.ec.gc.ca/climate_normals. [Accessed June 2009].

Gendron, F., Messier, C., and Comeau, P.G. 1998. Comparison ofvarious methods for estimating the mean growing season percentphotosynthetic photon flux density in forests. Agric. For. Me-teorol. 92(1): 55–70. doi:10.1016/S0168-1923(98)00082-3.

George, L.O., and Bazzaz, F.A. 2003. The herbaceous layer as afilter determining spatial pattern in forest tree regeneration. InThe herbaceous layer in forests of eastern North America. Edi-ted by F.S. Gilliam and M.R. Roberts. Oxford University Press,New York. pp. 265–282.

Gilliam, F.S. 2007. The ecological significance of the herbaceouslayer in temperate forest ecosystems. Bioscience, 57(10): 845–858. doi:10.1641/B571007.

Gilliam, F.S., and Roberts, M.R. 2003. Interactions between theherbaceous layer and overstory canopy of eastern forests: a me-chanism for linkage. In The herbaceous layer in forests of east-ern North America. Edited by F.S. Gilliam and M.R. Roberts.Oxford University Press, New York. pp. 198–223.

Goldberg, D.E., Rajaniemi, T., Gurevitch, J., and Stewart-Oaten, A.1999. Empirical approaches to quantifying interaction intensity:competition and facilitation along productivity gradients (Meta-Analysis in Ecology). Ecology, 80(4): 1118–1131. doi:10.1890/0012-9658(1999)080[1118:EATQII]2.0.CO;2.

Grime, J.P. 1977. Evidence for the existence of three primary stra-tegies in plants and its relevance to ecological and evolutionarytheory. Am. Nat. 111(982): 1169–1194. doi:10.1086/283244.

Grime, J.P. 2001. Plant strategies, vegetation processes, and eco-system properties. John Wiley & Sons Ltd., Chichester.

Hart, S.A., and Chen, H.Y.H. 2006. Understory vegetation dy-namics of North American boreal forests. Crit. Rev. Plant Sci.25(4): 381–397. doi:10.1080/07352680600819286.

Hautala, H., Tolvanen, A., and Nuortila, C. 2007. Recovery of pris-tine boreal forest floor community after selective removal of un-derstorey, ground and humus layers. Plant Ecol. 194(2): 273–282. doi:10.1007/s11258-007-9290-0.

Hicks, S., and Turkington, R. 2000. Compensatory growth of three

herbaceous perennial species: the effects of clipping and nutrientavailability. Can. J. Bot. 78(6): 759–767. doi:10.1139/cjb-78-6-759.

Howard, J.L. 1993. Linnaea borealis. In Fire Effects InformationSystem [online]. U.S. Department of Agriculture, Forest Service,Rocky Mountain Research Station, Fire Sciences Laboratory(Producer). Available from www.fs.fed.us/database/feis/ [Ac-cessed February 2010].

Jiang, L., Wan, S., and Li, L. 2009. Species diversity and produc-tivity: why do results of diversity-manipulation experiments dif-fer from natural patterns? J. Ecol. 97(4): 603–608. doi:10.1111/j.1365-2745.2009.01503.x.

Kishchuk, B. 2004. Soils of the Ecosystem Management EmulatingNatural Disturbance (EMEND) experimental area, northwesternAlberta. Natural Resources Canada, Canadian Forest Service,Northern Forestry Centre, Edmonton, Canada. Inf. Rep. NOR-X-397.

Kohyama, T., and Takada, T. 2009. The stratification theory forplant coexistence promoted by one-sided competition. J. Ecol.97(3): 463–471. doi:10.1111/j.1365-2745.2009.01490.x.

Kotliar, N.B., and Wiens, J.A. 1990. Multiple scales of patchinessand patch structure: A hierarchical framework for the study ofheterogeneity. Oikos, 59(2): 253–260. doi:10.2307/3545542.

Kun, A., and Oborny, B. 2003. Survival and competition of clonalplant populations in spatially and temporally heterogeneous ha-bitats. Community Ecol. 4(1): 1–20. doi:10.1556/ComEc.4.2003.1.1.

Legendre, P., and Anderson, M.J. 1999. Distance-based redundancyanalysis: testing multispecies responses in multifactorial ecologi-cal experiments. Ecol. Monogr. 69(1): 1–24. doi:10.1890/0012-9615(1999)069[0001:DBRATM]2.0.CO;2.

Leniere, A., and Houle, G. 2009. Short-term response of the un-derstory to the removal of plant functional groups in the cold-temperate deciduous forest. Plant Ecol. 201(1): 235–245.doi:10.1007/s11258-008-9545-4.

Lieffers, V.J., Macdonald, S.E., and Hogg, E.H. 1993. Ecology andcontrol strategies for Calamagrostis canadensis in boreal forestsites. Can. J. For. Res. 23(10): 2070–2077. doi:10.1139/x93-258.

Macdonald, S.E., and Fenniak, T.E. 2007. Understory plant com-munities of boreal mixedwood forests in western Canada: nat-ural patterns and response to variable-retention harvesting. For.Ecol. Manage. 242(1): 34–48. doi:10.1016/j.foreco.2007.01.029.

Maguire, D.A., and Forman, R.T.T. 1983. Herb cover effects ontree seedling patterns in a mature hemlock-hardwood forest.Ecology, 64(6): 1367–1380. doi:10.2307/1937491.

Magurran, A.E. 2004. Measuring biological diversity. BlackwellPublishing, Malden, Massachusetts.

McArdle, B.H., and Anderson, M.J. 2001. Fitting multivariate mod-els to community data: a comment on distance-based redun-dancy analysis. Ecology, 82(1): 290–297. doi:10.1890/0012-9658(2001)082[0290:FMMTCD]2.0.CO;2.

McCune, B., and Grace, J.B. 2002. Analysis of Ecological Commu-nities. MjM Software Design, Glenedon Beach, Oregon.

McCune, B., and Mefford, M.J. 1999. PC-ORD. Multivariate ana-lysis of ecological data. MJM Software Design, GlenedonBeach, Oregon.

Moss, E.H. 1983. Flora of Alberta. A manual of flowering plants,conifers, ferns and fern allies found growing without cultivationin the province of Alberta, Canada. Second edition. Revised byJ.G. Packer. University of Toronto Press.

Neufeld, H.S., and Young, D.R. 2003. Ecophysiology of the tempe-rate deciduous forests. In The herbaceous layer in forests ofeastern North America. Edited by F.S. Gilliam andM.R. Roberts. Oxford University Press, New York. pp. 38–90.

920 Botany Vol. 88, 2010


Nilsson, M.C., and Wardle, D.A. 2005. Understory vegetation as aforest ecosystem driver: evidence from the northern Swedishboreal forests. Front. Ecol. Environ, 3(8): 421–428. doi:10.1890/1540-9295(2005)003[0421:UVAAFE]2.0.CO;2.

Niva, M., Svensson, B.M., and Karlsson, P.S. 2006. Effects of lightand water availability on shoot dynamics of the stoloniferousplant Linnaea borealis. Ecoscience, 13(3): 318–323. doi:10.2980/i1195-6860-13-3-318.1.

Oborny, B., and Kun, A. 2002. Fragmentation of clones: how doesit influence dispersal and competitive ability? Evol. Ecol. 15(4–6): 319–346. doi:10.1023/A:1016084815108.

Oborny, B., Kun, A., Czaran, T., and Bokros, S. 2000. The effectof clonal integration on plant competition for mosaic habitatspace. Ecology, 81(12): 3291–3304. doi:10.1890/0012-9658(2000)081[3291:TEOCIO]2.0.CO;2.

Økland, R.H. 1995. Persistence of vascular plants in a Norwegianboreal coniferous forest. Ecography, 18(1): 3–14. doi:10.1111/j.1600-0587.1995.tb00114.x.

Oksanen, L., Sammul, M., and Magi, M. 2006. On the indices ofplant-plant competition and their pitfalls. Oikos, 112(1): 149–155. doi:10.1111/j.0030-1299.2006.13379.x.

Oksanen, J., Kindt, R., Legendre, P., O’Hara, B., Simpson, G.L.,Solymos, P., Henry, M., Stevens, H., and Wagner, H. 2009. Ve-gan: community ecology package. User’s guide and applicationpublished at: http://cc.oulu.fi/~jarioksa/.

Pennings, S.C., and Callaway, R.M. 2000. The advantages of clonalintegration under different ecological conditions: a community-wide test. Ecology, 81(3): 709–716. doi:10.1890/0012-9658(2000)081[0709:TAOCIU]2.0.CO;2.

Roberts, M.R. 2004. Response of the herbaceous layer to naturaldisturbance in North American forests. Can. J. Bot. 82(9):1273–1283. doi:10.1139/b04-091.

Roberts, M.R., and Gilliam, F.S. 2003. Response of the herbaceouslayer to disturbance in eastern forests. In The herbaceous layerin forests of eastern North America. Edited by F.S. Gilliam andM.R. Roberts. Oxford University Press, New York. pp. 302–320.

Rowe, J.S. 1956. Uses of undergrowth plant species in forestry.Ecology, 37(3): 461–473. doi:10.2307/1930168.

SAS Institute Inc. 2003. SAS. Version 9.1 [computer program].SAS Institute Inc. Cary, North Carolina.

Schwinning, S., and Weiner, J. 1998. Mechanisms determining the

degree of size asymmetry in competition among plants. Oecolo-gia (Berl.), 113(4): 447–455. doi:10.1007/s004420050397.

Smith, M.D., and Knapp, A.K. 2003. Dominant species maintainecosystem function with non-random species loss. Ecol. Lett.6(6): 509–517. doi:10.1046/j.1461-0248.2003.00454.x.

Strong, W.L., and Leggat, K.R. 1992. Ecoregions of Alberta. Al-berta Forestry, Lands and Wildlife, Land Information ServicesDivision, Resource Information Branch, Edmonton, Canada.Publication number T/245.

Stuefer, J.F., de Kroon, H., and During, H.J. 1996. Exploitation ofenvironmental heterogeneity by spatial division of labour in aclonal plant. Funct. Ecol. 10(3): 328–334. doi:10.2307/2390280.

Symstad, A.J., and Tilman, D. 2001. Diversity loss, recruitmentlimitation, and ecosystem functioning: lessons learned from a re-moval experiment. Oikos, 92(3): 424–435. doi:10.1034/j.1600-0706.2001.920304.x.

ter Braak, C.J., and Smilauer, P. 2002. CANOCO reference manualand Cano-Draw for Windows user’s guide: software for canoni-cal community ordination (version 4.5) [computer program]. Mi-crocomputer Power, New York.

Thiffault, N., Jobidon, R., De Blois, C., and Munson, A. 2000.Washing procedure for mixed bed ion exchange resin deconta-mination for in situ nutrient adsorption. Commun. Soil. Sci.Plant Anal. 31(3): 543–546. doi:10.1080/00103620009370456.

Thompson, A., Askew, A.P., Grime, J.P., Dunnett, N.P., and Willis,A.J. 2005. Biodiversity, ecosystem function and plant traits inmature and immature plant communities. J. Ecol. 19(2): 355–358. doi:10.1111/j.0269-8463.2005.00936.x.

Tilman, D. 2007. Interspecific competition and multispecies coexis-tence. In Theoretical ecology: principles and applications. Editedby R. May and A. McLean. Oxford University Press, Oxford.pp. 84–97.

Wardle, D.A., and Zackrisson, O. 2005. Effects of species andfunctional group loss on island ecosystem properties. Nature,435(7043): 806–810. doi:10.1038/nature03611. PMID:15944702.

Appendix A

Table A1. List of vascular plants found in the sample plots(Nomenclature follows Moss 1983).

Tree speciesAbies balsameaBetula papyriferaPicea glaucaPopulus balsamiferaPopulus tremuloides

Shrub speciesAlnus crispaAmelanchier alnifoliaCornus stoloniferaLedum groenlandicumLonicera dioicaRibes lacustreRibes oxyacanthoidesRosa acicularisRubus idaeusSalix spp.Shepherdia canadensisVaccinium caespitosumVaccinium myrtilloidesVaccinium vitis-idaea



Viburnum edule

Herb speciesGraminoidsCalamagrostis canadensisElymus innovatus

ForbsAchillea millefoliumActaea rubraAnemone canadensisArnica cordifoliaAralia nudicaulisAstragalus americanusAster ciliolatusAster conspicuusCalypso bulbosaCircaea alpinaCornus canadensisDelphinium glaucumEpilobium angustifoliumEquisetum arvenseEquisetum pratenseEquisetum scirpoidesEquisetum sylvaticumFragaria virginianaGalium borealeGalium triflorumGeocaulon lividumGoodyera repensHabenaria obtusataHabenaria orbiculataLathyrus ochroleucusLinnaea borealisLycopodium annotinumLycopodium complanatumMaianthemum canadenseMertensia paniculataMitella nudaMoneses unifloraOsmorhiza depauperataPetasites palmatusPyrola asarifoliaPyrola secundaPyrola virensRubus pubescensViola canadensisViola renifolia

922 Botany Vol. 88, 2010


understory species interactions in mature boreal mixedwood forests

Documents