Botanical Studies (2007) 48: 339-348.

*

Corresponding author: E-mail: Zhangjt@bnu.edu.cn

INTRODUCTION

Mountain forest communities perform an array of

important ecosystem functions, including water and soil

conservation (Wang, 1991; Molles, 2002), the provision

of important animal and wildlife habitat, maintenance of

biological diversity (Wu, 1982; Sparks, 1995), and the

development of ecotourism (Cheng and Zhang, 2003).

In North China, most natural forests can only be found

in mountainous areas. The virgin forests of North China

occur only in the Lishan National Nature Reserve (LNNR),

Shanxi province (Jiang, 1986). The conservation of

mountain forest communities in this reserve is significant

and of wide interest (Liu, 1984; Fu and Zheng 1994;

Zhang et al., 1997; Zhang, 2003).

In China, the need for proper characterization of

natural forests is important in light of the mandates of

the National Conservation Project for Natural Forests of

1999 (State Forestry Administration, 1999) and recent

efforts to pursue ecosystem management, both of which

require the use of effective quantitative approaches to

ensure that management practices maintain the integrity

and biodiversity of forest ecosystems (Zhang, 2002).

Many studies of mountain forests have used multivariate

statistical techniques to characterize vegetation patterns

(Zhang, 1995; Zhang et al., 1997; Loreau et al., 2001;

Leps and Smilauer, 2003; Zhang and Chen, 2004), but

few studies have attempted to incorporate other vegetation

layers in the evaluation of vegetation patterns and

underlying environmental gradients using quantitative

methods (Mi and Zhang, 1995; Lyon and Sagers, 2002).

One approach to addressing the complexity of

mountain forests is functional analysis. Plant species can

be classified into functional groups based on a variety

of characteristics. Each functional group potentially will

partition the environmental gradient differently (Smith

and Huston, 1989; Austin, 1990; Dale, 1998; Lyon and

Sagers, 2002). Thus, spatial and temporal changes in

environmental resources will also affect functional groups.

eCOlOgy

Diversity and composition of plant functional groups in

mountain forests of the lishan Nature Reserve, North

China

Jin-Tun ZHANG

1,

* and Feng ZHANG

2

1

College of Life Sciences, Beijing Normal University, Beijing 100875, P.R. China

2

Institute of Botany, Chinese Academy of Science, Beijing 100093, P.R. China

(Received October 12, 2006; Accepted February 16, 2007)

ABSTRACT.

Canonical correspondence analysis (CCA) was used to characterize the composition and

distribution of forest vegetation within the Lishan National Nature Reserve (LNNR), Shanxi Province, China.

The LNNR is located at E111o05¡¦43"-111o56¡¦29", N35o29¡¦07"-35o23¡¦10", and is part of the Zhongtiao

mountain range. Forest vegetation was sampled from 58, 10 m ¡Ñ 20 m plots along an elevation gradient

from 1,400 to 2,100 m. Floristic and environmental data of different functional groups, such as trees, shrubs,

saplings and herbs, were analyzed using CCA; and the changes of species richness, diversity and evenness of

different functional groups were analyzed in relation to environmental variables. Overall, all the functional

groups of forest vegetation showed a statistically significant correlation with elevation and soil Cu. Responses

to other environmental gradients differed among the four groups of plants analyzed. Tree layer showed a

correlation with soil P, shrubs and herbs showed a correlation with soil organic matter and N, while saplings

showed a correlation with slope and aspect. Elevation was the most important variable in terms of variations

in species diversity. Species richness, evenness, and diversity of different functional groups showed a similar

responding model to changes in elevation, i.e. the maximum diversity occurring at intermediate elevations.

Of the functional groups analyzed, trees were most important in maintaining species evenness within

communities, while shrubs and herbs were significant in maintaining species richness within communities of

the LNNR.

Keywords: CCA; Functional groups; Ordination; Soil variables; Species diversity; Topographic factors;

Vegetation-environment relationship.

340

Botanical Studies, Vol. 48, 2007

The use of vertically stratified growth forms or vegetation

layers as a means of separating species functional groups

has a sound ecological and physiological basis (Grime,

1993; Box, 1996). However, information concerning

the differential responses of different forest layers to

environmental gradients is limited in the literature

(Pausas, 1994; Zhang, 2003; Lyon and Sagers, 2002).

The hypothesis of species diversity-elevation gradient has

been tested many times (e.g. Stevens, 1992; Dolezal and

Srutek, 2002), but information concerning the responses

of functional group species diversity to environmental

gradient is limited (Lomolino, 2001; Kessler, 2001;

Austrheim, 2002).

The LNNR was established in the 1970s to protect its

virgin forest communities. Vegetation within the reserve

is typical of warm temperate deciduous broadleaved

forest in China. Previous studies have examined the flora

(Liu, 1984), vertical distribution of vegetation (Zhang

et al., 1997), and plant resources (Liu, 1984; Fu and

Zheng, 1994) of the reserve. However, the relationships

between forest functional groups and environments have

not been studied. The objectives of this study were to

evaluate the roles of species functional groups in the

forest communities, determine if different functional

groups exhibit different responses to the same suite of

environmental variables, and determine the patterns of

species diversity in different functional groups.

MATeRIAlS AND MeTHODS

Study area

The LNNR is located at E111o05¡¦43"-111o56¡¦29",

N35o29¡¦07"-35o23¡¦10", and is part of the Zhongtiao

mountain range in Southern Shanxi, China. Elevation

within the reserve varies from 1,000 to 2,358 m. The

reserve lies on the southern edge of the Loess Plateau

and is within a transitional area from a warm-temperate

zone to a subtropical zone (Wu, 1982). The climate of the

area is warm temperate and semi-humid, with continental

characteristics and controlled by seasonal winds.

Annual mean temperature is 13.3

¢X

C, the monthly mean

temperatures of January and July are -0.5

¢X

C and 27.5

¢X

C,

respectively, and the annual accumulative temperature

in excess of 10

¢X

C is 2,100

¢X

C. Annual mean precipitation

varies from 667.6 mm to 900.0 mm in the mountain

regions, with 70% of annual precipitation during July and

September. The dominant soil types are drab (cinnamon)

soil, mountain drab soil, brown forest soil and mountain

meadow soil according to Chinese soil classification

system (Liu, 1992). The study area of Zhuweigou is the

main valley within the LNNR. Its elevation varies from

1,400 to 2,358 m. Most of the valley is forested, while

a small area close to the mountain tops is covered by

mountain shrub-lands and meadows. This study dealt with

all forests distributed from 1,400 to 2,100 m elevation.

The vegetation communities are mainly broad-leaved

deciduous forests (Zhang et al., 1997).

Vegetation sampling

We established 8 eight transects within Zhuweigou

valley at intervals of 100 m in altitude along an elevation

gradient from 1,400 to 2,100 m. These transects cut across

the valley and were oriented parallel to topographic

contours. Four to eight sites along each transect were

established randomly; the number of sites along each

transect was determined based on the transect length.

The plot size was 10 m ¡Ñ 20 m, based on the minimum

community area of 128 m

2

in this area (Mi and Zhang,

1995). A total of 58, 10 m ¡Ñ 20 m plots were sampled

during our survey in July, 2002.

We categorized plants as trees (>2 cm dbh, height

5-15 m), tree saplings (<2 cm dbh, height 2-5 m), shrubs

(height <2 m) and herbs (height <0.5 m). Correspondingly,

we segregated the vegetation into four a priori defined

functional groups: tree layer species, tree sapling species,

shrub layer species, and herb species. The criteria for

division into functional groups were the height layers of

plants; tree saplings and shrubs were separated because

saplings are generally much taller than shrubs and have

special roles in forest regeneration, and adult trees and

saplings for the same species may have different roles in

the forest (Grime, 1993; Loreau et al., 2001). The cover,

height, diameter at 1.3 m height (dbh), basal area, and

individual number for tree species were measured in each

10 m ¡Ñ 20 m plot. Three 4 m ¡Ñ 4 m and four 2 m ¡Ñ 2 m

subplots within each 10 m ¡Ñ 20 m plot were used to record

shrubs and herbs, respectively; the cover and height for

shrubs and tree saplings were measured in each 4 m ¡Ñ 4

m subplot, and that for herbs was measured in each 2 m ¡Ñ

2 m subplot. We calculated the mean values of cover for

shrub, sapling and herb in the three 4 m ¡Ñ 4 m or four 2 m

¡Ñ 2 m subplots, and the mean values of heights for species

in the subplots. The cover of plants was estimated visually,

and the heights were measured using tree-finder for trees

and saplings, and tape-ruler for shrubs and herbs. The db

h diameters and basal diameters of trees were measured

using calipers and were used to calculate stem and basal

areas.

environmental attributes

Each study plot was characterized by three physical

factors: elevation, slope, and aspect. The elevation for

each plot was measured by altimeter, while the slope and

aspect were measured by compass. The 8 classes of aspect

of 1 (337.6¢X-22.5¢X), 2 (22.6¢X-67.5¢X), 3 (292.6¢X-337.5¢X),

4 (67.6¢X-112.5¢X), 5 (247.6¢X-292.5¢X), 6 (112.6¢X-157.5¢X),

7 (202.6¢X-247.5¢X), 8 (157.6¢X-202.5¢X) were used in the

analysis (Zhang, 2004). We collected soil samples at 20

cm depth from five locations chosen randomly within each

10 m ¡Ñ 20 m plot using a soil cylindered core sampler. The

five samples were thoroughly mixed and one quarter of the

mixture was collected for laboratory chemical analysis.

Soil samples were dried at 70

¢X

C for over 48 h to a constant

weight and analyzed in terms of soil pH, conductivity,

organic matter, total nitrogen, total phosphorus, K, Cu,

ZHANG and ZHANG ¡X Forest plant functional groups in Lishan Reserve

341

Mn, and Zn. These variables were selected because some,

such as N, P, K, and organic matter, are the most important

nutrient elements, while others, such as the micronutrient

elements Cu, Mn, and Zn, are deficient in soils of the

area (Liu, 1992). A 1:2.5 ratio of soil to distilled water

suspension was used to measure pH and conductivity using

a Whatman pH sensor meter and a conductivity meter,

respectively. Total nitrogen was estimated using Kjeldahl

extraction, and total phosphorus was measured via the

HCLO

4

-H

2

SO

4

colorimetric method (molybdovanadate

method). Organic matter was measured using K

2

Cr

2

O

7

-

capacitance. The elements K, Cu, Mn, Zn were determined

using an atomic absorption spectrophotometer (Page,

1982).

Data analysis

We used the importance value (IV) of each species as

data in ordination analysis and the calculation of diversity

indices. The IV was calculated using the following

formulae (Zhang, 1995, 2004):

IV

Tree

= Relative cover + Relative dominance + Relative

height

[1]

IV

Tree sapling

,

Shrub or Herbs

=

Relative cover + Relative height

[2]

where the dominance refers to the sum of the basal areas

for each tree species within a plot; relative cover, relative

dominance and relative height refer to the percentages

of one species cover, dominance and mean height over

the sum of all species cover, dominance and mean height

within a plot respectively.

Canonical correspondence analysis (CCA) was

conducted on plant species-environmental variable

matrices using the software CANOCO 4.5 (ter Braak

and .milauer, 2002). The square root transformation

of environmental data was used, but no transformation

of species data (IVs) was applied in the analysis. To

determine if different functional groups exhibited

differential responses to the same suit of environmental

variables, separate CCA ordinations were performed on

the four functional groups, i.e. tree layer, sapling layer,

shrub layer and herb layer.

To determine the patterns of species diversity, we

employed three species diversity indices: species richness,

species diversity (diversity), and species evenness:

Species number (as a richness index):

D

= S.

[3]

Shannon-Wiener diversity index:

H ¡¦ = -

.

P

i

ln P

i

.

[4]

Pielou evenness index:

E = (-

.

P

i

ln P

i

)/lnS.

[5]

where P

i

is the relative importance value of species i,

and S is the number of species within a plot (Pielou, 1975;

Zhang, 1995; Tothmeresz, 1995). Calculation of species

diversity indices was performed for the four functional

groups.

Linear and non-linear (quadratic curves) regression

methods were used to analyze the relationships between

the CCA axes of the four functional groups, and between

species diversity indices and environmental variables. The

regression analyses were performed using SPSS (SPSS

Inc. 2000), and the regression diagrams were plotted using

Microsoft Excel.

ReSUlTS

Ordination analysis of functional groups

We performed CCA ordinations on trees, shrubs, tree

saplings and herbs respectively. In all CCA ordinations,

the Monte Carlo permutation test indicated that the

eigenvalues for the first four axes were all significant

(P<0.05; Ter Braak, 1986). The species-environment

correlations with the CCA axes for all functional groups

were significant (Table 1), however, the relationships

between species and environmental variables differed for

different plant groups (Figures 1-4, Table 1, Table 2).

Nine of the twelve environmental variables are

significantly correlated with tree species distribution

(Figure 1, Table 2). The dominant environmental variables

correlated with the first axis were soil Cu and elevation

(Figure 1, Table 2). Elevation, P and Mn showed a strong

correlation with the second axis, while slope and aspect

showed a correlation with the third axis. The plots at

high elevation that contain Betula albo-sinensis, Populus

davidiana, Salix pseudotangii, an d Pinus armandii are

located in the upper left quadrant of the CCA triplot

(Figure 1). Species that are characteristic of low-elevation

plots, such as Pinus tabulaeformis, Juglans cathayensis,

Table 1. Comparison of eigenvalues and species-environment correlations produced by CCA ordinations on the four functional

groups in the Lishan Nature Reserve, China.

Eigenvalues

Species-environment correlations

Axis 1 Axis 2 Axis 3 Axis 4

Axis 1 Axis 2 Axis 3 Axis 4

Trees

0.784 0.470 0.394 0.235

0.925 0.837 0.834 0.82

Scrubs

0.644 0.481 0.356 0.233

0.962 0.866 0.893 0.791

Tree sapling

0.838 0.602 0.506 0.440

0.946 0.841 0.798 0.810

Herbs

0.535 0.371 0.323 0.282

0.928 0.858 0.843 0.829

342

Botanical Studies, Vol. 48, 2007

Carpinus turczaninowii var. stipulata, Carpinus

turczaninowii, plot in the right-hand quadrants. Species

that prefer fertile soil and that occur in mid-elevation

plots, such as Acer davidii, Quercus liaotungensis, Acer

mono and Toxicodendron verniciflum , plot in the lower left

quadrant.

As with the CCA ordination of trees, shrubs showed

a strong correlation between the first axis and elevation

and soil Cu (Table 2, Figure 2). Soil organic matter, N and

pH were also important to the first axis. The dominant

environmental variables that correlate with the second

axis were P and N, while the dominant variables that

correlate with the third axis were slope and aspect. The

high-elevation plots contained Abelia biflora, Spiraea

pubescens, Philadelphus incanus and Lonicera chrysantha

plot in the upper left quadrant of the CCA triplot (Figure

2). Species characteristic of low-elevation plots, such

as Vitex negundo var. heterophylla and Rosa xantheana

plot in the upper right quadrant, while species that prefer

fertile soil and that occur in mid-elevation plots, such as

Philadelphus incanus, Forsythia suspensa and Sambucus

williamsii, plot in the lower quadrants.

For tree saplings, the dominant variable correlated

with the first axis was slope, while aspect and elevation

were also significantly correlated with the first axis

(Table 2, Figure 3). The dominant variables correlated

with the second and third axes were slope and elevation

respectively. Cu was also significantly correlated with the

first three axes. Saplings were unevenly distributed in the

study area; 18 plots do not have any saplings.

Herbaceous species showed strong correlations

between the first axis and elevation, Cu, organic matter

and N (Table 2, Figure 4), similar to the trees and shrubs

ordinations. Zn, pH and conductivity were also important

to the first axis. Zn, pH, and P showed a significant

correlation with the second axis, while aspect and slope

showed a strong correlation with the third axis. The

Figure 1. CCA ordination of 58 plots and 18 tree species (> 2

cm dbh) with 12 environmental variables in the Lishan Nature

Rese rve, China. Biplot vectors s hown represe nt the major

explanatory environmental variables. Ele: Elevation; Slo: slope;

Asp: aspect; Org: organic matter content; Cond: conductivity. ¡´

and regular font numbers represent plots, ¡³ and italic numbers

represent tree species. 1, Carpinus turczaninowii; 2, Carpinus

turczaninowii var. stipulata; 3, Ulmus lamellosa; 4, Acer davidii;

5, Sorbus pohuashanensis; 6, Quercus liaotungensis; 7, Juglans

cathayensis; 8, Betula platyphylla; 9, Betula albo-sinensis; 10,

Pinus armandii; 11, Pinus tabulaeformis; 12, Salix pseudotangii;

13, Populus davidiana; 14, Toxicodendron verniciflum; 15, Acer

mono; 16, Rhus chinensis; 17, Tilia mongolica; 18, Quercus

aliena.

Figure 2. CCA ordination of 58 plots and 55 shrub s pecies

with 12 environmental variables in the Lishan Nature Reserve,

China. Biplot vectors shown represent the major explanatory

environmental variables. Ele: Elevation; Slo: slope; Asp: aspect;

Org: organic matter content; Cond: conductivity. ¡´ and regular

font numbers represent plots, ¡³ and italic numbers repres ent

shrub species. 1, Staphylea holocarpa; 2, Swida alba; 3, Cerasus

polytricha; 4, Euodia rutaecarpa; 5, Crataegus kans uensis;

6, Cornus bretschneideri; 7 , Koelreuteria paniculata; 8,

Syringa reticulata var. mandshur ica; 9, Pyrus betulaefolia;

10, Malus honanens is; 11, Hy dr angea brets chneider i; 12,

Celastrus orbiculatus; 13, For sythia suspensa; 14, Lonicera

chrysantha; 15, Lonicera hispida; 16, Lonicera ferdinandii;

17, Lonicera microphylla; 18, Spiraea pubescens; 19, Spiraea

trilobata; 20, Vitex negundo var. heterophylla; 21, Sambucus

williamsii; 22, Abelia biflora; 23, Cotoneaster multiflorus; 24,

Rubus crataegifolius; 25, Lespedeza bicolor; 26, Schisandra

chinensis; 27, Cotoneaster acutifolius; 28, Sorbaria sorbifolia;

29, Philadelphus incanus ; 30, Acanthopanax gracilistylus ;

31, Acanthopanax senticosus; 32, Euonymus alatus; 33,

Euonymus nanoides; 34, Ribes mandshuricum; 35, Viburnum

betulifolium; 36, Viburnum schensianum; 37, Viburnum opulus

var. calv esc ens; 38, Smilax china; 39, Ros a xanthina; 40,

Rosa davurica; 41, Rosa bella; 42, Hydrangea bretschneideri;

43, Ber ber is amurensis ; 44, Syr inga microphylla; 45, Ribes

bruejense; 46, Pyrus betulaefolia; 47, Myr ipnois dioica; 48,

Daphne odora; 49, Sophora flavescens ; 50, Rhamnus davurica;

51, Elaegnus pungens; 52, Celastrus orbiculatus; 53, Elsholtzia

stauntoni; 54, Crataefus pinnatifida; 55, Diospyros lotus.

ZHANG and ZHANG ¡X Forest plant functional groups in Lishan Reserve

343

high-elevation plots that contain species Thalictrum

squarrosum, Polygonum convolvulus, Cimicifuga foetida,

Dryopteris laeta and Maianthemum bifolium plot in the

left quadrants. Plots with N-rich soils were found in the

lower left quadrant, while plots with Zn-rich soils were

located in the upper left quadrants. The low-elevation

plots contained the species Patrinis jeterophylla, Carex

lanceolata, Phlomis umbrosa, Artemisia lavandulaefolia

and Dryopteris laeta plots were in the right quadrants;

their soils were rich in Cu with comparatively high pH.

Species diversity of functional groups

Elevation was one of the most significant variables that

correlate with vegetation and species functional groups

from the CCA ordination analyses (Table 2, Figures 1-4).

Therefore, we examined variations in species diversities

of functional groups along the elevation gradient (Figure

5). Species richness, diversity and evenness of the three

main functional groups, trees, shrubs and herbs, varied

throughout the study area (diversity of saplings was not

considered separately because saplings were not found in

more than 30% plots). Species richness of trees and herbs

Figure 3. CCA ordination of 58 plots and saplings of 20 tree

s pecies (< 2 cm dbh) with 12 environmental variables in the

Lishan Nature Reserve, China. Biplot vectors shown represent

the major explanatory environmental variables. Ele: Elevation;

S lo: s lope; Asp: aspect; Org: organic matter content; Cond:

conductivity. ¡´ and regular font numbers represent plots, ¡³ and

italic numbers represent sapling species. 1, Pinus tabulaeformis;

2, Ulmus lamellose; 3, Koelreuter ia paniculata; 4, Juglans

cathayensis; 5, Cornus bre tschneider i; 6, Pinus armandii;

7, Carpinus turcz aninowii var. stipulata; 8, Toxicodendron

verniciflum; 9, Quercus liaotungensis; 10, Acer davidii; 11, Acer

mono; 12, Carpinus turczaninowii; 13, Ailanthus altiss ima;

14, Sorbus pohuashanensis; 15, Tilia mongolica; 16, Fraxinus

chinensis; 17, Populus davidiana; 18, Syringa reticulata var.

mandshurica; 19, Salix pseudotangii; 20, Swida alba.

Figure 4. CCA ordination of 58 plots and 85 herbaceous species

with 12 environmental variables in the Lishan Nature Reserve,

China. Biplot vectors shown represent the major explanatory

environmental variables. Ele: Elevation; Slo: slope; Asp: aspect;

Org: organic matter content; Cond: conductivity. ¡´ and regular

font numbers represent plots, ¡³ and italic numbers represent herb

species. 1, Smilacina japonica; 2, Phlomis umbrosa; 3, Veratrum

nigrum; 4, P ar is ver ticillata; 5, Triosteum pinnatifidum; 6,

Astilbe chinensis; 7, Cacalia hastata; 8, Cystopteris fragilis; 9,

Dryopteris laeta; 10, Aquilegia viridiflora; 11, Vicia unijuga;

12, Patrinia heterophylla; 13, Dioscorea nipponica; 14,

Polygonatum odoratum; 15, Gal ium v erum; 16, Ar isae ma

erubescens; 17, Epimedium grandiflorum; 18 , Saus s urea

japonica var. alata; 19, Aconitum carmichaeli; 20, Aconitum

barbatum; 21, Ajuga ciliata; 22, Ligularia intermedia; 23, Viola

prionantha; 24, Viola varietata; 25, Viola biflora; 26, Artemisia

lavandulaefolia; 27, Artemisia argyi; 28, Artemisia brachyloba;

29, Pedicularis resupinata; 30, Thalictrum petaloideum;

31, Thalictrum squarrosum; 32, Achyr anthes bidentata; 33,

Rubia cordi folia; 34, Ser ratul a chinens is; 35, P impinella

thellungiana; 36, Hylotelephium verticillatum; 37, Polygonatum

involucratum; 38, Pertya sinensis; 39, Polygonum convolvulus;

40, Lamium album; 41, Carpesium cernuum; 42, Glycine soja;

43, Chrysosplenium pilosum; 44, Polygonatum verticillatum;

45, Lactuca tatarica; 46, Lonicera tragophylla; 47, Kalimeris

lautureana; 48, Her acleum hemsleyanum; 49, Maianthemum

bifolium; 50, Cimicifuga foetida; 51, Sedum aizoon; 52, Pyrrosia

lingua; 53, Les pedeza davurica; 54, Cardamne tangutorum;

55, Aster tataricus; 56 , Angelica dahurica; 5 7, Paeonia

obovata; 58, Convallar ia keiske i; 59, Hyperic um erec tum;

60, Cirsium teo; 61, Poa annua; 62, Arundinella hirta; 63,

Dendranthena lavandulifolium; 64, Polygonum viviparum; 65,

Cynanchum ascyrifolium; 66, Allium senescens; 67, Potentilla

chinensis; 68, Leontopodium leontopodioides; 69, Car ex

lanceolata var. subpeditormis; 70, Carex rigescens; 71, Carex

arnellii; 72, Viola yedoensis; 73, Iris tenuifolia; 74, Anemone

tomentosa; 75, Leibnitzia anandria; 76, Cleistogenes serotina;

77, Geum urbanum; 78, Stellaria media; 79, Dendranthema

cavandulifolium; 80, Vitis amurensis; 81, Ixeris sonchifolia; 82,

Duchesnea indica; 83, Akebia quinata; 84, Cacalia farfaraefolia;

85, Clematis macropetala.

344

Botanical Studies, Vol. 48, 2007

showed a significant correlation with elevation, but the

relationship between shrub richness and elevation was not

significant (Figure 5). Tree richness showed a quadratic

increase with increasing elevation and reached a maximum

value at 1,900-2,000 m. In contrast, herb richness

decreased with increasing elevation, and reached a lowest

value at 1,750-1,800 m, before it increased again at higher

elevations. The species richness rank of functional groups

is herbs > shrubs > trees (Figure 5).

The species heterogeneities (diversities) (Shannon-

Wiener index) of functional groups showed a distinct

correlation with elevation (Figure 5). Heterogeneities of

trees, shrubs and herbs all showed a quadratic curve: the

value increased at low elevations to a maximum value

before decreasing at high elevation. The diversity curve of

shrubs reached a maximum value at 1,850 m (Figure 5).

Similar to richness, the species diversity rank of functional

groups is herbs > shrubs > trees.

The relationship between species evenness of

functional groups and elevation is unimodal with a peak at

intermediate elevations (Figure 5). The evenness of trees,

shrubs and herbs reached maximum values at elevations of

1,850 m, 1,800 m and 1,900 m, respectively. The variation

of tree evenness was smaller than that of shrubs and herbs

in the study area. The species evenness rank of functional

groups is trees > shrubs > herbs, which is different to that

of species richness and diversity.

DISCUSSION

A variety of plant characteristics, including

morphology, physiology, reproductive, competitive

status or location in a successional sere could be used to

define functional groups (Grime, 1993; Zhang, 2005).

We separated forest plants into four functional layers,

trees, saplings, shrubs and herbs, based on their height,

which reflected importance of community structure

(Zhang, 1999; Lyon and Sagers, 2002). The heights

v ari ed fr om 5 m t o 1 5 m, 2 m to 5 m , and < 2 m f or

trees, saplings and shrubs respectively in LNNR, and the

heights of herbaceous species were usually below 0.5 m.

Correspondingly, the depth of their root system varied

greatly (Begon et al., 1990; Molles, 2002; Zhang, 2003).

Therefore, the responses of these functional layers to

environmental variables may differ and should be studied.

This work demonstrated the distinctive responses of

different functional groups to environmental gradients and

confirmed the necessity of this kind of study.

Forest vegetation and their functional groups in

the LNNR responded directly or indirectly to several

important variables, including elevation, soil Cu, soil

organic matter, soil N, slope and aspect. The responses of

woody vegetation to these gradients have been observed

in other studies (Nigh et al., 1985; Lyon and Sagers,

1998; Verburg and van Eijk-Bos, 2003; Chen et al., 2004),

however, these previous studies usually focused on the

distribution of canopy species within a forest. Lyon and

ZHANG and ZHANG ¡X Forest plant functional groups in Lishan Reserve

345

Zhang et al., 1997) and therefore their effects on vegetation

and species were significant. These variables were more

significant to shrubs and herbs (Table 1-2, Figure 2 and

Figure 4), because these variables were richer in shallow

soil than in deep soil, and the root systems of shrubs and

herbs were shallower than those of trees (Nangendo et al.,

2002). Variations in slope were obvious, and its effects on

saplings were apparent because it affected the maintenance

and germination of seeds in soils. For instance, the number

of seeds of Quercus variabilis within soil decreases with

increasing slope in the Zhongtiao Mountains (Fu and

Zheng, 1994; Zhang, 2004). Steep slopes make it difficult

for seeds to stay at the soil surface, to move within the

soil, and to absorb sufficient water (Zhang et al., 1997).

These relationships must be considered when investigating

the regeneration of forest in the LNNR.

Although differences in response to environmental

gradients among functional groups were obvious, their

similarities were clear from the CCA ordination results.

Figure 6 shows the correlations of the first two CCA axes

of one functional group with that of the other. All the axes

show a significant correlation with other axes, similar

to other studies (Franklin et al., 2001; Lyon and Sagers,

2002; Nangendo et al., 2002). Different functional groups

Sagers (2002) investigated the responses of all trees and

shrubs to environmental gradients, and identified specific

relationships between environmental variables and woody

vegetation. A more detailed approach, incorporating all

species groups (trees, shrubs, saplings and herbs), reveals

some of the specific interactions between environment and

forest communities.

The differences in responses to environmental gradients

between functional groups were clear in the LNNR

forests. Elevation and soil Cu are dominant variables in

affecting trees, shrubs and herbs, whereas slope is most

important to tree saplings. In addition to elevation and

Cu, soil P is important to trees, while organic matter and

N are significant to shrubs and herbs. Herbs also show

close relationship to pH and Zn. The relationship between

elevation and vegetation in the LNNR has been noted in

previous studies (Liu, 1984; Fu and Zheng, 1994; Zhang

et al., 1997). The Cu content of soils of the LNNR is high

compared with other mountain areas, but Cu distribution

is uneven in the study area (Liu, 1992); some communities

were Cu-deficient. Therefore, Cu is significant to forest

communities in the LNNR (Wu, 1982). Organic matter, N

and P are basic nutrient resources for plants, however, their

concentrations vary markedly with elevation (Liu, 1984;

Figure 5. Variations of species diversity of functional groups along elevational gradient in the Lishan Nature Reserve, China.

346

Botanical Studies, Vol. 48, 2007

in a community share the same complex of environmental

variables, and therefore, it is easy to understand that they

are similar in their responses to environmental gradients.

This indicates that the functions and roles of species

groups in communities are closely related to each other

(Whitaker, 1967; Zhang, 2003; Nummelin and Zilihona,

2004). Although different plant groups show similar

patterns of species diversity in response to elevation, those

elevations with the maximum diversity values are different

from each plant group (Figure 5); this is due to interactions

between functional groups (Figure 5). The effects of trees

on the diversity of shrubs and herbs are significant because

tree canopies affect the distribution of resources such as

light, water-conditions and temperature available to shrubs

and herbs (Kessler, 2001; Zhang, 2003; Nummelin and

Zilihona, 2004). Among trees, shrubs and herbs, trees

were most important in maintaining species evenness

of communities, and shrubs and herbs were significant

in maintaining species richness of communities in the

LNNR (Zhang and Chen, 2004). All the functional groups

were important in maintaining species diversity (Loreau

et al., 2001; Verburg and van Eijk-Bos, 2003), because

species diversity is related to both species richness and

evenness (Zhang, 1995). To protect species diversity and

forests, the conservation of all functional layers should be

emphasized. Elevation gradient is one key variable that

affects the variation of species diversity in communities

and is frequently studied (Lyon and Sagers, 2002; Zhang,

2002). The patterns of species diversity of functional

groups along the elevation gradient were very similar; the

species richness, diversity and evenness of different plant

groups all show a significant correlation with elevation,

with most being unimodal, consistent with the hypothesis

of maximum diversity at intermediate elevations

(Lomolino, 2001; Zhang and Chen, 2004). One exception

was the richness of herbs, which had a minimum richness

at intermediate elevations.

The distribution patterns of vegetation and species

diversity are often correlated with patterns of resource

variation and resource gradients, which have been well

established in vegetation science (Whittaker, 1967; Austin,

1990; Zhang, 2002). Different plant functional groups

may have different resource-use strategies, physiologies,

and competitive abilities (Lyon and Sagers, 2002). In the

literature, many vegetation studies avoid complexity by

over-emphasizing a single type of plant group, typically

a group of canopy tree species (O¡¦Neill et al., 1986). In

the present study, an analysis of all species, including

canopy trees, understory shrubs, saplings and herbs,

emphasized the importance of the structure of complex

forest ecosystems. The inclusion of different functional

groups in multivariate analysis provides more detailed

and comprehensive information on the spatial variation of

vegetation and species diversity and the response of plant

species to underlying gradients (Lyon and Sagers, 2002;

Zhang, 2004). Furthermore, vegetation functional group

analyses can be used in a management and conservation

context to identify vegetation and landscape characteristics

(Lyon and Sagers, 2002; Verburg and van Eijk-Bos, 2003).

Acknowledgment. We thank G.L. Zhang, Q. Li and

X.Y. Guo for their help with data collection. The study

was financial supported by the National Natural Science

Foundation of China (Grant No 30070140) and the

Teachers¡¦ Foundation of Education Ministry of China.

Figure 6. Relationships of the first two CCA ordination axes of one functional group with the other functional group in the Lishan

Nature Reserve, China.

ZHANG and ZHANG ¡X Forest plant functional groups in Lishan Reserve

347

lITeRATURe CITeD

Austin, M. P. 1990. Community theory and competition in

vegetation. In J. B. Grace and D. Tilman (eds.), Perspectives

on Plant Competition. Academic P ress, New York, pp.

215-238.

Aus trheim, G. 2002. Plant diversity patterns in semi-natural

gra ss la nds a long a n ele vati onal gradie nt in s out hern

Norway. Plant Ecol. 161: 193-205.

Begon, M., L . Harper, and C.R. Towns end. 1990. Ecology.

Blackwell Scientific Publications, Boston, pp. 267-350.

Box, E.O. 1996. Plant functional types and climate at the global

scale. J. Veg. Sci. 7: 309-320.

Chen, J., B. Song, M. Rudnicki, M. Moeur, K. Bible, M. North,

D.C. Shaw, J.F. Franklin, and D.M. Braum. 2004. Spatial

relationship of biomass and species distribution on an old-

growth Pseudotsuga-Tsuga forest. Forest Sci. 50: 364-375.

Cheng, Z.H. and J-T. Zhang. 2003. Difference between tourism

vegetation landscapes of different distance. J. Mountain Sci.

21: 647-652.

Dale, G. 1998. Forest plant diversity at local and landscape

s c a le s i n th e Ca s c ad e Mo unt a in s o f S o ut hw es t e rn

Washington. Forest Ecol. Manag. 109: 323-341.

Dolezal, J. and M. Srutek. 2002. Altitudinal changes in

c om p os i ti on a nd s t ruc t ur e of m oun ta i n- te m pe ra t e

vegetation: A case s tudy from the Western Carpathians.

Plant Ecol. 158: 201-221.

F ranklin, J., A.D. Syphard, D.J. Mla denoff, H.S . He, D.K.

Simons , R.P. Martin, D. Deutschman, and J .F. O¡¦Leary.

2001. Simulating the effects of different fire regimes on

plant functional groups in Southern California. Ecol. Model.

142: 261-283.

F u, Z.J. and X.T. Zheng. 1994. The floristic feature of plants

in Zhongtiao Mountains, Shanxi. Acta Bot. Boreali-

Occidentalia Sin. 14: 148-152.

Grime, J.P. 1993. Vegetation functional classification systems as

approaches to predicting and quantifying global vegetation

change. In A.M. S ol om on a nd H .H. S hug art (e ds .),

Vegetation Dynamics and Global Change. Chapman and

Hall, New York, pp. 293-305.

J iang, S .Z . 1986. The vi rgin fores ts i n Hungou dis trict of

Zhongtiao Mountains. J. Beijing Forest. Univ. 8: 67-70

Kessler, M. 2001. Patterns of diversity and range size of selected

plant groups along an elevational transect in the Bolivian

Andes. Biodivers. Conserv. 10: 1897-1921.

Leps, J. and P. Smilauer. 2003. Multivariate Analysis of

Ecological Data Using CANOCO. Cambridge University

Press: Cambridge.

L iu, T.W. 1984. A survey of plant re source s in Z hongt iao

Mountains of Shanxi. J. Wuhan Bot. Res. 2: 259-266.

Liu, Z.Y. 1992. Soils in Shanxi Province. Science Press, Beijing.

(in Chinese).

Lomolino, M.V. 2001. Elevation gradients of species diversity:

historical and prospective views. Global Ecol. Biogeogra.

10: 3-13.

Loreau, M., S . Naeem, and P. Inchausti. 2001. Biodivers ity

and ecosystem functioning: current knowledge and future

challenges. Science 294: 804-808.

Lyon, J. and C.L. Sagers. 1998. Structure of herbaceous plant

assemblages in a forested riparian landscape. Plant Ecol.

138: 1-16.

Lyon, J. and C.L. Sagers. 2002. Corres pondence analys is of

functional groups in a riparian landscape. Plant Ecol. 164:

171-183.

Mi, X-C. and J-T. Zhang. 1995. Classification and ordination of

Form. Quercus variabilis in Menghe Reservation, Shanxi

Province. Bull. Bot. 15: 397-402.

Molles, M.C. 2002. Ecology: Concepts and Applications (2

nd

ed). The McGraw-Hill Co., Singapore.

Nangendo, G., A. Stein, M. Gelens, A. de Gier, and R. Albricht.

2002. Quantifying differences in biodivers ity between

a tropic al fore st area and a g ras sl and are a s ubjec t to

traditional burning. Forest Ecol. Manag. 164: 109-120.

Nigh, T.A., S.G. Pallardy, and H.E. Garrett. 1985. Sugar maple-

environment relationships in the river hills and central

Ozark Mountains of Missouri. The Amer. Midl. Naturalist

98: 469-476.

Nummelin, M. and I.J.E. Zilihona. 2004. S patial variation of

arthropod communities in virgin and managed sites in the

Kibale Forest, western Uganda. Forest Ecol. Manag. 195:

107-114.

O¡¦Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen.

1986. A hierarchical concept of ecosystems. Monographs

i n P opula tion Biology 23. P rinceton University P re ss,

Princeton, New York.

Page, A.L. (ed). 1982. Methods of soil analysis, chemical and

m icrobiological properties , Part 2. Madison: Am erican

Society of Agronomy, pp. 831-871.

Pausas, J.G. 1994. Species richness patterns in the understorey

of Pyrenean Pinus sylvestris forest. J. Veg. Sci. 5: 517-524.

Pielou, E.C. 1975. Ecological Diversity. London: Wiley & Sons.

Smith, T. and M. Hus ton. 1989. A theory of the spatial and

temporal dynamics of plant communities. Vegetatio 83:

49-69.

Sparks, R.E. 1995. Need for ecosystem management of large

rivers and their floodplains. BioScience 45: 168-182.

SPSS Inc. 2000. SPSS software. Headquarters, 233 S. Wacker

Drive, 11th floor, Chicago, Illinois, USA.

State Forestry Administration. 1999. The Key Projects of Natural

F orest Protection. The Forestry P ress House, Beijing. (in

Chinese)

Stevens, G.C. 1992. The elevational gradient in altitudinal range:

an extension of Rapoport¡¦s latitudinal rule to altitude. Amer.

Naturalist 140: 893-911.

Ter Braak, C.J.F. 1986. Canonical correspondence analysis: a

new eigenvector technique for multivariate direct gradient

analysis. Ecology 67: 1167-1179.

348

Botanical Studies, Vol. 48, 2007

Ter Braak, C.J.F. and P. .milauer. 2002. CANOCO Reference

Manual and User¡¦s Guide to Canoco for Windows. Software

f or Can onic al C omm uni ty Ordi nat ion (ve rs ion 4.5 ).

Centre for Biometry Wageningen (Wageningen, NL) and

Microcomputer Power (Ithaca NY, USA), 352 pp.

Tothmeresz, B. 1995. Comparison of different methods for

diversity ordering. J. Veg. Sci. 6: 283-290.

Verburg, R. and C. van Eijk-Bos. 2003. Effects of selective

logging on tree diversity, composition and plant functional

type patterns in a Borne an rain fores t. J. Veg. S ci. 14:

99-110.

Wang, YF. 1991. Vegetation Resources and Their Utilization and

Protection in Loess Plateau. China Science and Technology

Press, Beijing. (in Chinese)

Whittaker, R.H. 1967. Gradient analysis of vegetation. Biol.

Rev. 42: 207-264.

Wu, Z.Y. 1982. Vegetation of China. Science Press, Beijing, pp.

453-615. (in Chinese)

Zhang, J-T. 1995. Methods in Quantitative Vegetation Ecology.

China Science and Technology Press: Beijing. (in Chinese)

Zhang, J-T. 1999. Species abundance patterns in some broad-

le a ve d d ec i duo us f ore s ts in Ne w Yo rk, US A. Ac ta

Phytoecol. Sin. 23: 481-489.

Zhang, J-T. 2002. A study on relations of vegetation, climate and

soils in Shanxi province, China. Plant Ecol. 162: 23-31.

Zhang, J-T. 2003. Applied Ecology. Science Press: Beijing. (in

Chinese)

Zhang, J-T. 2004. Quantitative Ecology. Science Press: Beijing.

(in Chinese)

Zhang, J.-T. 2005. Succession analysis of plant communities in

abandoned croplands in the Eastern Loess Plateau of China.

J. Arid Environ. 63(2): 458-474.

Zhang, J-T. and T.G. Chen. 2004. Variation of plant communities

along an elevation gradient in the Guandi Mountains, North

China. Community Ecol. 5: 227-233.

Zhang, J-T., F. Zhang, and T.L. Shangguan. 1997. Re-analysis

of vertical vegetation spectral of Zhongtiao Mountains. J.

Shanxi Unv. 20: 76-79.