Does a shift in shade tolerance as suggested by seedling morphology explain differences in regeneration success of northern red oak in native and introduced ranges?

Peter Nosko·Kerri Moreau·Christian Kuehne·Kelly C.Major ·Jürgen Bauhus

Abstract Across North America,forests dominated by Quercus rubra L.(northern red oak),a moderately shadetolerant tree species,are undergoing successional replacement by shade-tolerant competitors.Under closed canopies,Q.rubra seedlings are unable to compete with these shadetolerant species and do not recruit to upper forest strata.In Europe,natural regeneration of introduced Q.rubra is often successful despite the absence of fire,which promotes regeneration in the native range.Considering that understorey light availability is a major factor affecting recruitment of seedlings,we hypothesized that Q.rubra seedlings are more shade tolerant in the introduced range than in the native range.Morphological traits and biomass allocation patterns of seedlings indicative of shade tolerance were compared for Q.rubra and three co-occurring native species in two closed-canopy forests in the native range (Ontario,Canada) and introduced range (Baden-Württemburg,Germany).In the native range,Q.rubra allocated a greater proportion of biomass to roots,while in the introduced range,growth and allocation patterns favored the development of leaves.Q.rubra seedlings had greater annual increases in height,diameter and biomass in the introduced range.Q.rubra seedlings in the introduced range were also younger;however,they had a mean area per leaf and a total leaf area per seedling that were five times greater than seedlings in the native range.Such differences in morphological traits and allocation patterns support the hypothesis that Q.rubra expresses greater shade tolerance in the introduced range,and that natural regeneration of Q.rubra is not as limited by shade as in the native range.The ability of Q.rubra seedlings to grow faster under closed canopies in Europe may explain the discrepancy in regeneration success of this species in native and introduced ranges.Future research should confirm findings of this study over a greater geographical range in native and introduced ecosystems,and examine the genetic and environmental bases of observed differences in plant traits.

Keywords Alien species·Cross-continental study·Forest regeneration·Introduced species·Seedling morphology·Plant functional traits·Quercus rubra

Introduction

Historically,oaks have been one of the most dominant and valued hardwood species in eastern North America (Abrams 1992;McShea et al.2007) and Europe (M?lder et al.2019;Kohler et al.2020).Due to their wide variety of ecological functions and services,members of this genus have been classified as both“keystone”and“foundation”species (Hanberry and Nowacki 2016;Mitchell et al.2019).Quercus rubraL.(northern red oak) is a mid-successional and moderately shade-tolerant species broadly distributed throughout the eastern deciduous forests of North America.In addition to its substantial ecological value (e.g.,high wildlife food and habitat quality),Q.rubrais one of the most important oaks harvested for lumber in its native range (Sander 1990)and has become one of the most valuable non-native species in Europe (Nicolescu et al 2018;Pettenkofer et al.2019).

As with other upland oaks in its native range,Q.rubrashows poor regeneration and has undergone successional replacement such that its relative abundance in mixed hardwood forests has steadily declined,while that of mesophytic shade-tolerant trees such asAcer rubrumL.(red maple)andAcer saccharumMarsh.(sugar maple) has increased(McDonald et al.2002;Aldrich et al.2005).The failure of oak to regenerate is linked to a variety of contributing factors,including climate change,seed predation,insect defoliation,ungulate herbivory,nitrogen deposition and timber harvesting (Lorimer 1993;Collins and Carson 2003;McEwan et al.2011;Frelich et al.2017).However,low subcanopy light levels and poor competitive performance under deep shade appear to be especially important contributing factors to failed regeneration and decline in the abundance ofQ.rubra(Crow 1988;Abrams 1992;Lorimer 1993;Brose 2008).

The success of natural regeneration ofQ.rubrastands is regulated by environmental factors that influence understorey light availability,growth characteristics associated with light capture in the face of intense competition,and life-history traits that persist in times of resource deficiency(Dey and Parker 1997;Alexander et al.2008;Jagodziński et al.2018).According to the Fire-Oak Hypothesis (Abrams 1992;Abrams and Nowacki 1992;Lorimer 1993;Brose et al.2013;Nowacki and Abrams 2008;Arthur et al.2012),successful regeneration ofQ.rubrarequires occasional disturbance such as that imposed by fire,which opens the canopy and controls competing species in the understorey.This favours fire-tolerantQ.rubraand creates the moderate light conditions that provides this species with an advantage over its more shade-tolerant but fire-sensitive competitors(Abrams 1 992;Abrams and Nowacki 1992).In forests historically dominated byQ.rubra,decades of active fire suppression have maintained closed canopies,allowing maple competitors for example,to suppressQ.rubraregeneration in shaded understoreys and midstoreys.Maples subsequently replaceQ.rubrain the canopy,as gaps become available(Abrams 1992).In the absence of fire,successful regeneration of native oak forests requires intensive management to maintain understorey light levels that are sufficient to allow recruitment of oak seedlings (Loftis 1990;Dey and Parker 1997).

In contrast with the decline ofQ.rubraforests in North America,populations of this species introduced to Europe in the late eighteenth century regenerate and expand naturally without fire disturbances (Tinner et al.2005) and with little or no silvicultural manipulation (Steiner et al.1993;Magni Diaz 2004;Vansteenkiste et al.2005;Chmura and Sierke 2006).Furthermore,in Europe,shade-tolerant trainer species such asCarpinus betulusL.(European hornbeam) andTiliaspp.(linden,lime or basswood) have often been planted withQ.rubrato suppress epicormic branching (Saha et al.2012;Kuehne et al.2013).In some European countries,Q.rubrais considered invasive (Reinhardt et al.2003;Vansteenkiste et al.2005;Nicolescu et al.2018;Chmura 2020)and concern has been expressed that this alien species is displacing native tree species,including indigenous oaks such asQuercus petraea(Matt.) Liebl.(sessile oak) andQ.roburL.(pedunculate oak) (Riep?as and Straigyt? 2008;Kiedrzyński et al.2011;Woziwoda et al.2019;Dyderski et al.2020;Dyderski and Jagodziński 2021).However,in a few countries such as Germany,Q.rubrais not considered invasive (Nicolescu et al.2018) depending on the risk assessment approach used (Bindewald et al.2020,2021).

In North America,recruitment ofQ.rubraseedlings into midstorey layers is uncommon and mortality rates forQ.rubrasaplings tend to be high (Lorimer 1 984;Lorimer et al.1994;Collins and Carson 2004).However,such recruitment is common,for example,in Germany where this species can dominate sub-canopy strata and form pure thickets in canopy gaps (Vor 2005;Major et al.2013).This suggests that in Europe,Q.rubra:(1) has the potential to be an effective and sometimes aggressive competitor under suitable conditions,even in mid-canopy strata (Major et al.2013) (rarely observed in North America);and,(2) does not seem to suffer the same competitive disadvantage in shaded sub-canopies as in North America.

We sought to identify key factors responsible for consistently successful regeneration ofQ.rubrain Europe,when natural regeneration in its native range is frequently unsuccessful.Previous work suggests that under closed canopy forests,natural regeneration of European populations ofQ.rubrais less restricted by low light conditions than in the native range (Vor 2005;Major et al.2013;Kuehne et al.2014).Cross-continental studies of other deciduous tree species have shown greater shade tolerance for populations in exotic versus native ranges (Shouman et al.2017).If this is correct,regardless of the underlying reasons,dissimilar levels of shade tolerance should be reflected by differences in functional morphological traits and biomass allocation patterns between North American and European ecotypes.However,these previous studies ofQ.rubradid not directly compare populations from the native and introduced ranges.Various morphological and physiological plant traits (e.g.,relatively low values for leaf thickness and root to shoot ratios,and relatively high leaf size,specific leaf area,leaf area ratio and chlorophyll levels) are considered reliable indicators of shade tolerance (Valladares and Niinimets 2008) and are commonly used to represent relative shade tolerance in trees (Abrams and Kubiske 1990;Valladres et al.2002;Kim et al.2005).Similarly,biomass allocation patterns in tree seedlings are strongly related to relative shade tolerance and site differences in light availability including former,more favorable light conditions after thinning (Walters et al.1993;Canham et al.1996;Beaudet and Messier 1998;Wang et al.2006).Intraspecific comparisons of woody species have shown that,relative to their native range,plants in an exotic range can undergo functional shifts that confer greater shade tolerance and more rapid growth(Leishman et al.2014;Heberling et al.2016;Shouman et al.2017).Increasingly,researchers have used the framework of the Plant Economic Spectrum (e.g.,Abdala-Roberts et al.2018) to group plant functional traits along a continuum from acquisitive (functional traits promote rapid resource acquisition and growth) to conservative (functional traits promote resource conservation and slow growth) strategies to explain the success of alien plant species (Tecco et al.2010,2013;Dyderski and Jagodziński 2019).Such a framework might provide a useful approach to examine differences between native and introduced populations ofQ.rubrafrom the perspective of regeneration success.

Our objective was to determine whether morphological traits indicative of shade tolerance were better developed inQ.rubraseedlings at locations within the introduced range compared to those in the native range.The approach was two-fold;(1) compare morphology and biomass allocation patterns ofQ.rubraseedlings to those of native sympatric species at representative locations in both native and introduced ranges ofQ.rubra;and,(2) compare these traits directly forQ.rubraseedlings at locations within native and introduced ranges.It was hypothesized that a comparison of morphological traits will support the premise thatQ.rubraseedlings in Europe are less inhibited by shade than those in North America.More specifically,it was predicted thatQ.rubraseedlings in Europe would have enhanced leaf traits associated with the acquisition of above-ground resources(i.e.,light and carbon) compared to those in North America,thereby exhibiting shifts away from traits that are more conservative (native range) to those that are more acquisitive (introduced range).Greater shade tolerance (≈ carbon fixation under low light) in introduced populations could explain enhanced recruitment ofQ.rubraseedlings in the introduced range.Insight to the factors promoting successful regeneration ofQ.rubrain Europe could help develop solutions to the problem of poor oak regeneration in North America on the one side and controlling this alien species in regions where it poses a threat to native European plants on the other.

Materials and methods

Site description

Studies were conducted in mesic hardwood forests in central Ontario,Canada and Baden-Württemberg in southwestern Germany.In each region,two representative locations were selected in which:(1) multilayered forest canopies were dominated or co-dominated by matureQ.rubra;(2) there was a high degree of canopy closure (>90%) resulting in deep understorey shade;(3) understoreys had significant densities of healthyQ.rubraseedlings and those of cooccurring tree species;and,(4) deer browsing and insect herbivory were negligible.All study sites had acidic soils with comparable levels of moderate fertility (unpublished data).

In Ontario,study sites were located near Mattawa (46°10′ N,78° 31′ W) and at the Petawawa Research Forest near Chalk River (45° 59′ N,77° 59′ W),both locations being within the Great Lakes-St.Lawrence Forest Region.The regional climate is continental,with warm summers and cold winters.The mean annual temperature that includes both study sites ranges from 3.8 °C to 5.2 °C.January and July temperatures (coldest and warmest months) average?17.6 °C and 25.0 °C,respectively,with approximately 184 days per year with temperatures above 10 °C (Environment Canada 2013).Mean annual precipitation is 898 mm with 703 mm falling as rain,mostly over the growing season.The Ontario sites are near the southern edge of the Precambrian Shield with bedrock consisting of granites and gneisses.Soils are well-drained and range from silty sands and sandy loams to fine sands and silt loams,often having well-decomposed and well-incorporated humus materials(OMNR 1998).The topography has been strongly influenced by glaciation and post-glacial outwashing with elevations varying from 140 to 280 m above sea level.The canopy at the Petawawa Research Forest was dominated byQ.rubraL.,Fagus grandifoliaEhrh.(American beech) andA.saccharum.The Klocks Road site near Mattawa had a forest canopy dominated byQ.rubraandA.rubrumwith lesser amounts ofA.saccharumandBetula alleghaniensisArnold(yellow birch).At this site,shade-tolerantOstrya virginiana(Mill.) K.Koch (ironwood) was prominent in the midstorey.The degree of vertical stratification of the Ontario stands was similar to the Baden-Württemberg sites,comprising a midstorey of saplings and young trees of shade tolerant species and dominant red oak in the overstorey.

The Baden-Württemberg sites were situated in the northern (48° 00′ N,07° 46′ E) and southern Mooswald (48° 01′N,07° 49′ E) near Freiburg im Breisgau.The study area is characterized by a warm Atlantic climate with a mean annual temperature of approximately 11 °C,mean January and July temperatures (coldest and warmest months)of 0.9 °C and 19.3 °C,respectively,and 185 days per year above 10 °C (Bl?sing 2008).A thirty-year average of growing degree days (GDD) above 10 °C for the period April 1 to September 30 shows that thermal input in Freiburg,Baden-Württemberg (1991 GDD) is greater than in Chalk River,Ontario (1429 GDD) (Weather Channel 2014).Mean annual precipitation is about 870 mm,with the majority(approximately 450 mm) occurring during the growing season between May and September (Gauer and Aldinger 2005).Situated in a former floodplain area of the Dreisam River,a gleyic cambisol (Michéli et al.2006) has developed from the alluvial deposits (silicate gravel),which are covered by a layer of loess of variable thickness (Villinger 2008).The well-drained soil is highly decarbonised,superficially acidic and has a gravely sandy texture.The terrain is flat with an elevation of approximately 220–230 m above sea level.The secondary and semi-natural forests of the northern and southern Mooswald sites comprise various hardwood species in the overstorey,includingQ.rubra,Acer pseudoplatanusL.(sycamore maple) andFraxinus excelsiorL.(European ash).C.betulusandTilia cordataMill.(smallleaved lime) form the mid-and understorey.

Study design and measurements

In Ontario,measurements forQ.rubraseedlings were compared to those of three shade-tolerant species,A.saccharum,A.rubrum,andF.grandifolia.In Baden-Württemberg,Q.rubrawas compared toQ.robur,an important indigenous oak having moderate shade tolerance,and two shade-tolerant species,A.pseudoplatanusandC.betulus.Measurements of environmental and seedling characteristics,as well as seedling harvests occurred in mid-August of 2009 and 2010,in Baden-Württemberg and in Ontario,respectively.

A representative 50 m×50 m plot was delineated at each study site,within which ten to twelve 2 m×2 m subplots were randomly located with the condition that overhead canopy closure was >90%.In these subplots,all seedlings of the four species of interest that ranged in height from 15 to 50 cm were tagged.In each region,20 (10 from each of the two study sites) whole seedlings per species were randomly selected for measurement and harvesting from among the tagged individuals.Subplots did not always contain eligible representatives of all study species.In a few cases,additional seedlings located just outside of the subplots had to be used to maintain a balanced study design.

From the centre of each sub-plot,mean canopy closure was determined by averaging measurements made in the four cardinal directions with a spherical densiometer (Lemmon 1956).Prior to harvest,the shoot height and collar diameter were recorded for each tagged seedling.Seedlings were selected that were healthy with no visible signs of herbivory and the rare seedlings that showed evidence of re-sprouting were rejected.Relative chlorophyll levels were measured for the top four leaves of each seedling (four readings per leaf)using a Minolta SPAD-502 chlorophyll meter.Seedlings of all species were carefully harvested in their entirety,including whole root systems.All leaves were removed from each seedling,dried and flattened in a plant press for the subsequent determination of leaf area (Epson 1640 XL scanner and Sigma Scan Pro 5 software),thickness (Mitutoyo digital thickness gauge) and dry weight.The remainder of each seedling was divided into stem,coarse root (>2 mm diam.)and fine root (<2 mm diam.) components.Seedling tissue components were oven-dried at 60 °C to a constant mass for at least 48 h prior to dry mass determination.The dry mass ratio for each seedling component was calculated relative to the total seedling biomass.Seedling ages were determined visually by counting annual rings at the stem base with the aid of a dissecting microscope and used to estimate annual height growth by dividing seedling height by age.

Statistical analyses

The study design does not allow the assumption that our samples are explicitly independent (Hurlbert 1984).However,many ecological field studies emphasize that achieving true replication can be difficult,logistically prohibitive or impossible (depending on the hypotheses),and that in the absence of true replication,valuable and ecologically significant insights can still be gained (Hargrove and Pickering 1992;Davies and Gray 2 015).Furthermore,in such cases,the use and presentation of inferential statistics has been encouraged by some researchers as a means to identify ecologically significant mechanisms (Oksanen 2001;Chaves 2010).Despite the constraints of our study design,we present our data alongside the results of statistical analyses with the understanding that the validity of our conclusions must be considered with caution (Davies and Gray 2015).

Following data exploration to determine whether assumptions of normality and homogeneity of variance could be met,differences in measured seedling traits forQ.rubraand sympatric species were determined for Ontario and Baden-Württemberg using one-way analyses of covariance(ANCOVA) with canopy closure and seedling age as covariates,using SPSS Statistics 18.0 (SPSS 2007).Resulting data were expressed as estimated marginal means (EMM).The degree to which ANCOVA considers the effect of covariates in adjusting the arithmetic mean to calculate the EMM is a function of the number of groups being compared,in our case,the number of species.Comparing,for example,OntarioQ.rubrato three sympatric species may give different EMM for OntarioQ.rubravariables than when comparing the same OntarioQ.rubradata to only the Baden-WürttembergQ.rubra.In all cases,differences were considered significant atp<0.05.

Results

Among the four Ontario species,no statistically significant differences were found in seedling height;annual height growth;stem diameter;annual radial stem growth;stem height-to-diameter ratio;total leaf area;the dry mass of total leaves,stems,total shoots or fine roots;as well as leaf-tostem ratio (Figs.1,2,3;Table S1).Q.rubraseedlings had the highest specific leaf area and lowest leaf mass and leaf area ratios (Fig.2;Table S1).Contrary to the expected trend,Q.rubrahad the highest level of leaf chlorophyll,being similar to that of the shade-tolerantF.grandifolia(Fig.2;Table S1).Compared to other Ontario species,Q.rubraseedlings allocated significantly more resources to root development,having the greatest dry mass for coarse roots and total roots,(neither of which were significantly different thanA.saccharum),as well as the greatest values for coarse root mass ratio and root-to-shoot ratio (not significantly different thanA.saccharum) (Fig.3;Table S1).

Fig.1 Canopy closure and seedling characteristics for Quercus rubra and sympatric tree species in Ontario (ON) and Baden-Württemberg (BW).Canopy closure and seedling age (mean+SD; n=20)were used as covariates in an ANCOVA.Other seedling traits are expressed as estimated marginal means (+SD; n=20) based on the two covariates.Results for testing of significant differences among species are in Tables S1 and S2.Abbreviations: Q.rub=Quercus rubra,A.rub= Acer rubrum,A.sac= Acer saccharum,F.gra= Fagus grandifolia,Q.rob = Quercus robur,A.pse=Acer pseudoplatanus and C.bet=Carpinus betulus

Fig.2 Leaf characteristics for seedlings of Quercus rubra and sympatric tree species in Ontario (ON) and Baden-Württemberg (BW)reported as estimated marginal means (+SD; n=20) based on the results of an ANCOVA using two covariates (canopy closure and seedling age).Results for testing of significant differences among species shown in Tables S1 and S2.Species abbreviations defined in Fig.1

Fig.3 Dry weights (a,b) and mass ratios (c,d) of seedling tissue components for Ontario(ON) (a,c) and Baden-Württemburg (BW) (b,d) tree species.Values are expressed as estimated marginal means(+SD; n =20) based on an ANCOVA using two covariates(canopy closure and seedling age).Results for testing of significant differences among species are shown in Tables S1 and S2.Abbreviations:L=leaf,ML=mean leaf,TL=total leaf,ST=stem,TSH=total shoot,FR=fine root,CR=coarse root,TR=total root,TSE=total seedling

In contrast with observations in Ontario,Q.rubrain Baden-Württemberg had the highest values for many of the measured variables compared to the indigenous tree species.For example,Q.rubraseedlings had significantly greater radial stem increment,total leaf area and dry mass of leaves(mean and total),total shoot,coarse roots,total roots,and total seedling than those of the three native Baden-Württemberg species (Figs.1,2,3; Table S2).Seedlings ofQ.rubraalso had the greatest annual height growth,stem collar diameter,mean leaf area and stem dry mass,although values for these variables were not always significantly greater than the other Baden-Württemberg species (Figs.1,2,3;Table S2).

In comparingQ.rubrain Ontario and Baden-Württemberg,the majority of seedling growth and biomass variables differed between the two locations (Figs.4,5).Chlorophyll levels forQ.rubraleaves were greater in Baden-Württemberg than in Ontario.In Baden-Württemberg,Q.rubraseedlings showed greater investment in photosynthetic tissue.Ontario seedlings had a significantly greater root-to-shoot ratio,suggesting the typical emphasis on the acquisition of belowground resources in the native range.Leaf-to-stem ratios indicated that a much greater proportion of aboveground biomass ofQ.rubraseedlings was allocated to photosynthetic (leaf) tissue and less to supportive (stem) tissue in Baden-Württemberg than in Ontario,suggesting a shift in the introduced range to enhanced light acquisition.

At any given age,Q.rubraseedlings in Baden-Württemberg consistently achieved greater height,diameter and total biomass than those in Ontario (Fig.6).The difference in seedling height and biomass between theseQ.rubraecotypes continued to become greater as seedlings aged (Fig.6).Young (<5 years)Q.rubraseedlings especially showed greater leaf area and leaf mass ratios in Baden-Württemberg than in Ontario;however,for progressively older seedlings,the differences diminished with these traits,becoming similar for nine year-old plants (Fig.6).

Fig.4 Seedling traits (estimated marginal meana based on ANCOVA using two covariates,±SD; n=20) of Quercus rubra in Ontario (ON)and Baden-Württemburg (BW).Differences between locations for covariates (reported as mean+SD) were examined using a one-way ANOVA.Statistical significance levels for trait differences between locations are indicated as NS=not significant (p >0.05),*=p≤0.05,**=p≤0.01,or ***=p≤0.001

Fig.5 Dry weights (a) and mass ratios (b) of tissue components for Quercus rubra seedlings in Ontario (ON) and Baden-Württemburg (BW).Values are expressed as estimated marginal means(+SD; n=20) based on an ANCOVA using two covariates (canopy closure and seedling age).Results for testing of significant differences among species are shown in Tables S1 and S2.Abbreviations:L=leaf,ML=mean leaf,TL=total leaf,ST=stem,TSH=total shoot,FR=fine root,CR=coarse root,TR=total root,TSE=total seedling.Statistical significance levels for trait differences between locations are indicated as NS=not significant (p >0.05),*=p≤0.05,**=p≤0.01,or ***=p≤0.001

Fig.6 Age related to a height,b collar diameter,c total biomass,d leaf area ratio,e leaf mass ratio,and f root-to-shoot ratio of Quercus rubra seedlings in Ontario (ON) and Baden–Württemberg (BW)(green).Trend lines fitted to scatterplots are based on logarithmic regression models with coefficients of determination (R2 ) shown

Discussion

Functional traits and resource acquisition of native and alien Q.rubra

Under similar canopy closure,growth and biomass allocation patterns forQ.rubraseedlings differed between Ontario and Baden-Württemberg sites in a manner consistent with different optimization strategies for acquiring belowground versus aboveground resources.As is characteristic ofQ.rubraacross its native range,growth trends were typical of a moderately shade-tolerant species with preferential allocation to root development,as evident from root-related variables that were consistently greater in Ontario than all other study species regardless of location.In Ontario,rootcentered growth reflects the stress-tolerant capabilities ofQ.rubrain dealing with shortages of belowground resources and explains the success of this species on dry,nutrient-poor sites (Kolb et al.1990;Johnson et al.2002;Rebbeck et al.2011).In some cases,leaf and shoot variables representing enhanced light capture were no different forQ.rubrafrom shade-tolerant species;however,such traits were usually most evident for the more shade-tolerantA.saccharumorF.grandifolia.

Atypically,height,diameter,and growth rates ofQ.rubrastems were similar to the more shade-tolerant Ontario species.For example,while seedlings ofF.grandifolia,the most shade-tolerant of the Ontario species (Kobe et al.1995;Beaudet and Messier 1998),did indeed have the greatest mean annual height growth,values did not differ significantly from those forQ.rubra.In shaded understoreys,Q.rubraseedlings have been characterized as having slow shoot growth that allows them to be easily over-topped by faster growing shade-tolerant species that are better able to maintain a positive carbon balance in low light (Lorimer 1993;Abrams 1998).Under such conditions,shade-tolerantA.saccharumseedlings,for example,can achieve relative growth rates that are 75% greater thanQ.rubra(Walters et al.1993).In our study,Q.rubrahad higher chlorophyll levels,and similar radial and vertical stem growth toA.saccharum;however,A.saccharumhad greater leaf area,thinner leaves and a lower specific leaf area.The conservative root-centred strategy ofQ.rubra,which allows persistence in stress-and disturbance-prone sites,could therefore be disadvantageous under current disturbance regimes in the mesophytic forests of eastern North America where the ability to compete for light is an important prerequisite for recruitment to the canopy.Seedling shoots that grow slowly are unlikely to overtake taller,shade-tolerant competitors once a critical height differential has been established under closed canopies (Crow 1988).

Among Baden-Württemberg species,dominance in shade-related functional traits ofQ.rubraseedlings was evident for many leaf variables.Despite significant investment in leaves,Q.rubraseedlings also maintained a high root biomass compared to native European species,so the tendency of early investment in roots persists in these introduced oak populations.In Baden-Württemberg,Q.rubraseedlings had a greater leaf mass ratio and formed larger leaves that had significantly greater biomass per leaf than in OntarioQ.rubra.Furthermore,more of the aboveground tissue ofQ.rubraseedlings was photosynthetic (leaves)rather than supportive (stems) in Baden-Württemberg compared to Ontario.Preferential investment in photosynthetic tissues ofQ.rubraseedlings is in keeping with the results of other comparative studies in Europe.For example,in Germany,enhanced leaf area production and highly efficient carbon acquisition inQ.rubraseedlings was reported under closed forest canopies compared with shade-tolerant European species (Kuehne et al.2014).In western Poland,Q.rubraseedlings not only had the greatest root mass fraction and total biomass,but also had the greatest total leaf area compared to seedlings of four native tree species (Dyderski and Jagodziński 2019).Studies ofA.pseudoplatanuswhich compared leaf functional traits in native (France) and introduced (New Zealand) ranges (Shouman et al.2017,2020)showed similar findings to our study.Just as we found forQ.rubra,A.pseudoplatanusexhibited more rapid growth and traits indicative of greater shade tolerance and competitive ability in the introduced range.Furthermore,as with our study of seedlings,Shouman et al.(2017) reported that leaf chlorophyll levels of trees (1–4 m.tall) were greater in the introduced range while in shaded sites,specific leaf area(SLA) in both ranges was similar.

One of the most frequently considered morphological indicators of an acquisitive or conservative strategy is specific leaf area (SLA);conservative plants have low SLA while acquisitive plants have high SLA (Tecco et al.2013;Weemstra et al.2016).Theory suggests that among woody plants,successful alien species should have more acquisitive traits (and presumably higher competitive ability) than native species (Tecco et al.2010).For example,A.pseudoplatanuswas reported to undergo a functional shift towards greater shade tolerance in its introduced range,allowing more rapid growth under low light and greater competitive ability,presumably due to an enhancement of acquisitive traits (Shouman et al.2017).Given the difference in regeneration success across native versus introduced ranges,Q.rubraseedlings could be assumed to have better-developed acquisitive traits at Baden-Württemberg compared to Ontario sites;however,mean SLA forQ.rubraseedlings was similar at both sites and to values reported by Dyderski and Jagodziński (2019).While Shouman et al.(2017)reported trait shifts forA.pseudoplatanustowards greater shade tolerance in the introduced range,SLA for native and introduced populations also remained the same.Dyderski and Jagodziński (2019) hypothesized that invasive alien tree species (includingQ.rubra) will have more acquisitive functional traits than their native competitors.They found thatQ.rubrabehaved differently than two other invasive tree species in retaining a conservative strategy (e.g.,low SLA,high root biomass) while achieving success by developing high individual biomass and leaf area,as found in our study.Our findings and those of Dyderski and Jagodziński (2019)suggest that in the introduced range,Q.rubragains no functional advantage via SLA but does so in terms of its biomass allocation strategy.This could be an adaptive feature of alienQ.rubrafacilitated by a high degree of variability in biomass traits,while plasticity in SLA is uncharacteristically low according to invasive theory (Dyderski and Jagodziński(2019).Interestingly,an interspecific study of oaks reported that,among 11 oak species,Q.rubraandQ.roburwere fully on the acquisitive side of the Plant Economic Spectrum(Abdala-Roberts et al.2018).

Implications for competitive ability and recruitment

The trends measured in this study are consistent with the premise thatQ.rubrais more shade tolerant and a more effective competitor in the regeneration layer of European compared to North American forests.This was also suggested in studies ofQ.rubrain Europe by Major et al.(2013) and Kuehne et al.(2014).The latter study,based on photosynthetic gas exchange,showed superior net carbon gain byQ.rubraseedlings at low and moderate light levels when compared to native species.Our data and those of Kuehne et al.(2014) suggest that this was facilitated by enhanced growth rates leading to large leaf areas further resulting in seedling dry masses that were greater than for native European species.This is supported byQ.rubraseedlings in Baden-Württemberg having twice the mean annual height increment and a relatively low mean root-to-shoot ratio compared to those in Ontario.Over approximately the first nine years of establishment,Q.rubraseedlings in our study showed greater leaf mass and leaf area ratios in Baden-Württemberg than in Ontario.Following this period,these leaf traits approached similarity in the native and introduced trees.However,stem heights and diameters,as well as the total biomass of Baden-WürttembergQ.rubraseedlings remained greater than in Ontario,suggesting that early investment in leaves translated to superior height growth and a lesser likelihood of being overtopped by competitors in Baden-Württemberg.AlthoughQ.rubraseedlings in Ontario and Baden-Württemberg did not differ significantly in height and total leaf area,Baden-Württemberg seedlings were younger (mean age=4.1 years) than in Ontario (mean age=7.2 years),supporting previous reports that in shaded forest understoreys,seedlings grow more rapidly in Baden-Württemberg (Major et al.2013) than in Ontario (Dech et al.2008).

Compared to their native range or to native plants in the introduced range,alien or exotic species often exhibit more rapid growth in the introduced range (Leishman et al.2014;Shouman et al.2017).In Baden-Württemberg,superior growth ofQ.rubrarelative to the native oak was apparent in that seedlings ofQ.rubraaccumulated almost twice the total biomass ofQ.roburdespite these seedlings not differing significantly in age or height.Furthermore,Q.rubrahad greater values for several variables related to leaf area,dry mass and mass ratios,including a greater proportion of photosynthetic versus supportive shoot tissues.While these trends suggest thatQ.rubrashould be a more effective competitor than the native oak,a similar fine root biomass suggests that belowground competition between the two oaks for water and nutrients might be less intense than aboveground competition for light under closed canopies,given the greater light acquisition capability ofQ.rubra.The larger leaves ofQ.rubraseedlings and their apparent superior competitive ability in Europe could explain the detrimental invasive effects that have been described forQ.rubraon local oak species such asQ.roburandQ.petraea(Riep?as and Straigyt? 2008;Kiedrzyński et al.2011;Chmura 2020).Woziwoda et al.(2019) compared natural regeneration of co-occurringQ.rubraandQ.roburin Europe.They concluded that advantages exhibited by the introduced oak did not arise from direct interspecific competition between seedlings or saplings;rather,the success ofQ.rubrawas attributed to a longer duration and frequency of acorn production,more effective colonization of empty niches and lower susceptibility to natural enemies.Several cross-continental studies have referenced the Enemy Release Hypothesis in emphasizing the importance of release from natural enemies.This is reinforced in a comparison showing that insect pests ofQ.rubraare more diverse in North America than in its introduced range (Dyderski et al.2020).With such release,greater resources are available to support rapid growth and enhance the competitive ability of successful/invasive introduced species (Adams et al.2009;Leishman et al.2014;Shouman et al.2017),especially when coupled with enhanced shade tolerance in the introduced range (Salgado-Luarte and Gianoli 2017).However,in a young tree diversity experiment near the Mooswald study sites in Freiburg,herbivory ofQ.rubraandQ.roburleaves did not differ (Wein et al.2016).

The assemblage of native species that interact with an invasive species can determine the conditions under which genetic adaptations are likely to enhance invasive success(Bossdorf et al.2005).The success ofQ.rubrain Baden-Württemberg could be influenced by a lesser degree of competitive pressure imposed by the sympatric native species compared to Ontario,as suggested when considering the effect of native competitor species onA.pseudoplatanusgrowing in native and introduced ranges (Shouman et al.2020).According to a meta-analysis of 806 species of woody plants from the temperate Northern Hemisphere (Niinemets and Valladares 2006),the shade tolerance ranking (most to least shade tolerant) of species used in our study isA.saccharum(Ontario) ≥F.grandifolia(Ontario) >C.betulus(Baden-Württemberg) >A.pseudoplatanus(Baden-Württemberg) >A.rubrum(Ontario) >Q.rubra(Ontario) >Q.robur(Baden-Württemberg).Accordingly,in shaded forest understoreys,the most highly shade-tolerant Ontario species should impose greater competitive pressure onQ.rubrathan the most shade-tolerant Baden-Württemberg species in our study.Non-nativeQ.rubrashould especially be compared withFagus sylvaticaL.(European beech),one of the most shade-tolerant hardwoods in Europe (Niinimets and Valladares 2006;Ellenberg 2009);however,F.sylvaticawas not a significant component of the forests in our study areas (Hügin 1990).In Europe,functional traits related to shade tolerance and acquisitive/conservative strategies ofQ.rubrawere compared to those ofF.sylvaticaby Dyderski and Jagodziński (2019) who reported that,as expected for the more shade tolerant species,F.sylvaticahad a lower root mass fraction,higher leaf mass fraction,higher SLA,and higher SLA plasticity.

Limitations and considerations for future research

Given the limitations of the study design,our findings must be considered with caution.Nevertheless,most elements of our findings are in close agreement with other studies ofQ.rubrain North America and Europe and are supported by cross-continental studies of various tree species.A focus is provided for future research to confirm our comparative findings over a greater expanse of sites in native and introduced ranges.Studies are also needed to disentangle genetic and environmental bases of observed differences in plant traits that influence the overlapping attributes of shade tolerance,competitive ability,regeneration success and invasive characteristics of this species.Intercontinental differences in the regenerative ability ofQ.rubracould be due to a combination of environmental,genetic or other biotic factors that differ between ranges.Compared to the numerous studies that have examined alien species in an introduced range (Major et al.2013;Kuehne et al.2014;Dyderski and Jagodziński 2019;Woziwoda et al.2019),few cross-continental studies have been conducted to compare a potentially invasive species in both introduced and native ranges (Hierro et al.2005;Adams et al.2009;Leishman et al.2014;Heberling et al.2016;Shouman et al.2017,2020).Such studies usually conclude by hypothesizing that range differences for environmental factors,phenotypic plasticity,shade tolerance,and susceptibility to pathogens and herbivores could be responsible for shifts in functional traits and resource-use strategies.However,as specific mechanisms remain unclear,researchers also conclude that reciprocal transplant or common garden studies are needed to disentangle the specific roles of these various factors (Bossdorf et al.2005;Heberling et al.2016;Shouman et al 2017).

With respect to the effect of environmental and genetic factors onQ.rubrain North America and Europe,intracontinental differences appear to be greater than those between continents (Merceron et al.2017).Notwithstanding the effect of intercontinental differences in genetic and environmental factors,functional traits indicative of shade tolerance tend to be conserved across climatic zones and forest types (Garnier et al.2001;Niinemets and Valladares 2006).Despite such variations,natural regeneration of red oak remains consistently impaired across all climate zones of its native range and consistently enhanced across Europe.For these reasons,we conjecture that trends observed in this study will be upheld across broader spatial scales in native and introduced ranges.

Conclusions

In shaded forest understoreys,Q.rubraseedlings showed significant differences in growth and biomass allocation patterns in Ontario and Baden-Württemberg consistent with differences in resource acquisition priorities,degree of shade tolerance,and potential for recruitment to upper forest strata.In Baden-Württemberg,Q.rubraseedlings showed greater investment in photosynthetic tissue,allowing for enhanced light and carbon acquisition,in contrast with the pattern of preferential investment in roots typical in its native range.Despite this early investment in leaves,exoticQ.rubraseedlings also maintain a high root biomass compared to native European tree species,which would not only promote efficient acquisition of belowground resources but also effectively support the development of photosynthetic tissue to a greater degree than observed in the native range.Our findings support the hypothesis thatQ.rubraexhibits greater shade tolerance in Baden-Württemberg than in Ontario.However,additional research is needed to confirm that our observed differences in growth and biomass allocation are evident across a greater portion of the native and introduced ranges ofQ.rubraand to examine the degree to which intercontinental differences in genetics,environmental factors,competitive pressure or forest management practices might be responsible.Such information could provide insights to effective management strategies for promotingQ.rubraregeneration in North America or to control its spread in Europe.The different patterns of growth,morphology and seedling biomass allocation observed in this study are likely key factors contributing to the intercontinental differences in natural regeneration success ofQ.rubra.

AcknowledgementsWe are grateful to Kevin Pigeau,Lisa Robinson,Lindsay Seguin and Marie-Cécile Gruselle for field assistance;Ursula Eggert,Anja Hausmann and Ernst Kraemer in Freiburg,and Peter Arbour at the Petawawa Research Forest and the Forestry Research Partnership in Ontario for logistic support.