Botanical Studies (2006) 47: 23-35.

*

Corresponding author: E- mail: ybgao@nankai.edu.cn; Tel:

+86-22-23508249; Fax: 86-22-23508800.

Morphological and RAPD analysis of the dominant

species

Stipa krylovii

Roshev. in Inner Mongolia steppe

Jin-Long WANG, Yu-Bao GAO*, Nian-Xi ZHAO, An-Zhi REN, Wei-Bin RUAN, Lei CHEN,

Jing-Ling LIU, and Chang-Lin LI

Department of Plant Biology and Ecology, College of Life Science, Nankai University, Tianjin 300071, P.R. China

(Received May 20, 2005; Accepted August 31, 2005)

ABSTRACT.

Stipa krylovii Roshev. is an important perennial tussock grass in the Inner Mongolian steppe.

It is found over a large area and has considerable ecological and economic importance. In the present study,

five natural populations of S. krylovii were selected from their typical habitats to study the quantitative trait

variation (samples from the natural populations) and RAPD variation. The relationships between quantitative

trait variation and RAPD variation, and between either variation and geographic distance and then climatic

factors were estimated using Mantel tests. Stipa krylovii populations showed differentiation in morphological

and RAPD characters, yet no significant relationship existed between genetic variations estimated by

morphological and RAPD characters and geographic distance, but both variations were closely associated

with the climatic variation. These results indicated that the populational differentiation of S. krylovii was not

in accord with the model of Isolation-by-distance, but was affected mainly by local climatic factors. Such

information would be useful for conservation managers working out an effective strategy to protect this

important species and provide the basis for a germplasm collection of it.

Keywords: Climatic factors; Genetic variation; Geographic distance; Mantel test; Quantitative traits; RAPD;

Stipa krylovii Roshev.

INTRODUCTION

Genetic variation is generally believed to be a

prerequisite for long- and short-term survival of a species

(Schonewald-Cox et al., 1983; Lande, 1988), and the

importance of preserving the genetic diversity of wild and

domesticated species is widely acknowledged today. Since

the development of bio-techniques in the 1960s, isozyme

and DNA molecular markers have been used frequently

to get variation estimates for plant species (Chung et al.,

1991; Kercher and Conner, 1996; Fahima et al., 1999;

Qian et al., 2001). The technique is relatively convenient,

yields a large number of useful markers, and often requires

very small amount of plant tissue (Fritsch and Rieseberg,

1996). Of these, the most popular marker is RAPD

because the technique is quick and reliable and therefore

enables a smooth evaluation of the molecular diversity in

a species (Black-Samuelsson et al., 1997).

Compared with traditional morphological analysis,

which may primarily indicate adaptation in a short as

well as long term perspective and can be performed

directly in a natural population or by quantitative genetic

studies of progenies under controlled conditions, the

molecular markers (isozyme and DNA markers) are

generally thought to be useful for detecting the action of

non-selective evolutionary forces, such as gene flow and

drift (Nei, 1987). Several examples illustrate that data

consisting only of selectively neutral markers may fail

to reveal adaptively important variation formed through

natural selection, and can lead to biologically unsound

management strategies (Olfelt et al., 2001). Therefore,

reports that quantitative genetic analyses served an

important complement in studies of plant species have

increased in the past several years. For example, using

isozymes and quantitative traits, Knapp and Rice (1998)

evaluated the patterns of genetic variation in Nassella

pulchra; Black-Samuelsson et al. (1997) analysed the

patterns of RAPD and morphological traits of the rare

plant species Vicia pisiformis. Olfelt et al. (2001) used

a combination of morphological and molecular genetic

markers to study the differentiation of Sedum integrifolium

in order to apply better conservation priorities and

design management strategies. These studies showed

that a combination of quantitative traits and molecular

markers to analyze the genetic patterns is powerful and

comprehensive.

Form. Stipa krylovii, consisting of all associations

dominated by the species itself, is one of the major

MOLECULAR BIOLOGY

24

Botanical Studies, Vol. 47, 2006

formations of the moderately-temperate steppe in central

Asia. The steppe is principally located in the inland

portion of the Eurasian continent. It is not only an

important pasture in Inner Mongolia, but also an important

green protective defence for the Beijing-Tianjin area,

and therefore has considerable economic and ecological

importance. Affected by a temperate continental climate,

the obvious thermal difference due to latitude and the

precipitation discrepancy due to monsoons provides the

possibility of genetic variation for S. krylovii in different

populations. Some populations of this grass might have

been genetically differentiated because the fragmented

habitats brought about by human activities could

have generated geographical isolation and population

differentiation driven by genetic drift (Baruch et al., 2004).

Zhao et al. (2003) studied the seed morphological traits

of eight different geographic populations of S. krylovii,

and the results showed that some significant differences

in awn length, seed length and seed diameter existed

among populations. Zhao et al. (2004) then analyzed the

genetic differentiation of seven S. krylovii populations

using RAPD markers. The results demonstrated significant

differentiation between populations at the DNA level and

that populations in similar habitats had closer genetic

distance values and could then be clustered into one

subgroup. However, studies that compare quantitative

traits and molecular markers of S. krylovii have not been

reported.

In the present study, five natural populations were

selected from the typical region of S. krylovii based on

their abundance, dominance, and representativeness of the

community. In this region, an environmental gradient in

aridity formed from the east to the west, and such gradient

provided an ideal model system for the study of genetic

variation by climatic variation. Accordingly, the purpose

of this study was to analyze patterns of morphological and

RAPD variations of S. krylovii in different locations and

climatic conditions, in order to examine whether climatic

factors were the main forces in modifying its genetic

structure, and whether two levels of population genetic

differentiation were affected by the same force. The

results will provide a general knowledge about population

genetic structure of S. krylovii. Such information will

be of great significance in explaining the distribution of

this important species and in developing conservation

strategies for it. The information will also provide the

basis for the germplasm collection of this species and for

the restoration of northern grassland in China, taking into

account that S. krylovii communities have been under the

pressure of over-grazing and urbanization over the last

several decades.

MATERIALS AND METHODS

Plant species

Stipa krylovii is a C

3

, tussock grass forming open

steppes that dominate the large semi-arid landscape of the

Inner Mongolian steppe, and the mature plant has very

dense tussocks about 20 cm high with long, thin leaves

and a very dense matrix of thin roots concentrated in the

first 20 cm of the soil layer. It is wind-pollinated, flowering

in middle or later July and ripening in late August or early

September. A seed consists of a callus, an awn, and a

lemma and palea encasing the caryopsis. The sharp callus

is covered with backward pointing hairs that readily attach

to animals, machinery, and clothing, thus aiding dispersal.

Lemmas are firm or even hardened and usually tightly

enclose the palea and caryopsis and with a long twisted

and twice-bent awn. As fruits dry, the awns become more

and more twisted, orienting the seed properly to the soil.

Study site

The plant materials used in this study were taken from

the five natural populations mentioned above. These

populations were located in the middle and eastern part

of the Inner Mongolian steppe in China (Figure 1). They

were from three different locations from the east to the

west, meadow steppe (one population), typical steppe (two

populations) and desert steppe (two populations). The five

populations were named Bayanwula, Xilinhot 1, Xilinhot

2, Xinhot, and Mandulatu after their locations from the

east to the west (Table 1). The Bayanwula population was

from the meadow steppe, where annual precipitation and

community species diversity were highest and cumulative

temperature in a year was lowest among the five sites. In

this community S. baicalensis and Filifolium sibiricum,

found only in meadow steppe, were dominant species, but

in relatively high hills, S. krylovii became dominant. The

Xilinhot 1 and Xilinhot 2 populations were located in the

east and west sides of Xilinhot City, respectively. Both

study sites were within the typical steppe where annual

precipitation was a little lower and cumulative temperature

in a year was higher than in meadow steppe, in which S.

krylovii and S. grandis were dominant species and the

species diversity was lower than that in Bayanwula. The

Figure 1. Sampling sites of S. krylovii populations (ƒÚ, ƒÛ, ƒÜ,

ƒÝ indicate the capital city of the corresponding administrative

regions, which are Bayanwula, Xilinhot, Xinhot and Mandulatu,

respectively).

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

25

Xinhot population was from the desert steppe, in which S.

krylovii was dominant species. The Mandulatu population

was from the most western site and within the desert

steppe, in which some typical desert species such as S.

gobica and Allium polyrhizum were found. In the latter

two sites, annual precipitation was lower than 250 mm,

and cumulative temperature in a year was rather higher

than at the other three sites.

Data collection for morphological characters

Data were collected for 20 quantitative traits (Table

2). The traits were chosen following previous work on

the genus Stipa (Zhao et al., 2003). In early September,

2004, 50 vegetative ramets, 50 reproductive ramets and

100 spikes were taken randomly in a quadrat of 50 ¡Ñ 50

m from each population site for morphological character

analysis. A distance of 3 m or more was left between

individual plants to ensure the independence of the

samples. A ramet with spike(s) on it was regarded as being

reproductive while one in a leafy stage was considered

vegetative. After a ramet (vegetative or reproductive) was

clipped above the ground, its actual length was measured

and labeled, and its dry weight was measured after being

air-dried at room temperature. A total of 100 spikelets

were collected from 100 spikes after being air-dried at

room temperature, and a set of measures¡Xincluding

diameter of seed, length of callus, length of lemma, length

of the first segment of awn, length of the second segment

of awn, length of awn apex, length of the first glume,

length of the second glume, length difference of the two

glumes, and length of awn¡Xwas obtained from an intact

spikelet.

RAPD analysis

In each of the five sites, eighteen individuals were

randomly collected at an interval of at least 10 m to

avoid collecting ramets from the same genet. Leaves

were harvested and stored with silica gel in zip-

Table 1. Location, soil type and habitat characters of the five S. krylovii populations.

Population Population code Administrative area Vegetation

Soil

Geographic position Altitude (m)

Bayanwula

POP1

Xiwu Banner Meadow steppe Dark Chestnut 44.64

¢X

N, 117.72

¢X

E 1152

Xilinhot 1

POP2

Xilinhot City Typical steppe Chestnut 44.14

¢X

N, 116.36

¢X

E 1121

Xilinhot 2

POP3

Xilinhot City Typical steppe Chestnut 43.93

¢X

N, 115.74

¢X

E 1088

Xinhot

POP4

Abaga Banner Desert steppe Light chestnut 44.12

¢X

N, 114.98

¢X

E 1267

Mandulatu

POP5

Dongsu Banner Desert steppe Light chestnut 43.83

¢X

N, 113.82

¢X

E 1157

Table 2. Description of morphological characters of S. krylovii plant.

Character

Brief description

Diameter of seed

The diameter of seed (with lemma) measured at the mid-point

Length of callus

Measured form callus tip to base of lemma

Length of lemma

Measured from base of lemma to base of awn

Length of the first segment of awn

Measured from the base of awn to the first geniculate point of awn

Length of the second segment of awn

Measured from the first to the second geniculate point of awn

Length of awn apex

Measured from the second geniculate point to the tip of awn

Length of the first glume

Measured from the base to the tip of the first glume

Length of the second glume

Measured from the base to the tip of the second glume

Length difference of the two glumes

The difference between the first glume length and the second glume length

Length of awn

Measured from the base to the tip of awn

Diameter of the second internode

The diameter of the second internode at its mid-point

Length of the second internode

Measured from node to the tip of the reproductive shoot

Length of flag leaf blade

Measured from the base to the tip of flag leaf blade

Height of reproductive shoot

Measured from top to bottom of the reproductive shoot

Dry matter weight per reproductive shoot

Height of vegetative shoot

Measured from top to bottom of the vegetative shoot

Leaf blade length per vegetative shoot

Sum of length of leaf blades of all leaves for a vegetative shoot

Dry matter weight per vegetative shoot

Length of longest leaf of vegetative shoot

Length of leaf sheath of first leaf

26

Botanical Studies, Vol. 47, 2006

lock bags for DNA extraction. Genomic DNA was

extracted using a modification of the protocol of

standard phenol-chloroform (Hillis et al., 1996). DNA

concentration and quality were determined with a UV-

VIS spectrophotometer (TU-1800) before it was diluted

to 30 ng/£gL and determined again in 0.7% regular agarose

(Spain) gels.

Eighty random decamer primers (hit A, I, N, Q, Operon

technologies, Inc., Alameda CA, USA) were tested

fo r PCR ampl ifi cat io n wi th tw o bu lk ed samp les. Th e

protocol for RAPD amplification described by Williams

(Williams et al., 1990) was optimized for use on a S.

krylovii template DNA. PCR was carried out in a 25 £gL

reaction volume containing about 30 ng template DNA

with 2.5 £gL 10¡Ñreaction buffer, 2.0 mM MgCl

2

, 0.2 mM

of each dNTP, 1 U Taq DNA polymerase and 0.2 mM

primer. The PCR reaction was run in a Programmable

Thermal Controller-100 (MJ research, Waltham MA,

USA), with the following temperature profile: preliminary

denaturation of DNA at 94¢XC for 4 min, followed by 40

cycles of 94¢XC for 1 min, 36¢XC for 1 min and 72¢XC for 2

min. After 40 cycles, there was a final step of 10 min at

72¢XC, followed by soaking at 4¢XC until recovery. RAPD

products were analyzed with electrophoresis in 1.5%

regular agarose (Spain) gels containing ethidium bromide

(0.5 mg/mL). Molecular marker SD005 (Beijing Dingguo

Biotechnology Development Center, China) was used as

a size marker. Gels were photographed under UV light

to ensure the RAPD reproducibility. The reproducibility

and repeatability of the amplification profiles were tested

for each primer. Only those bands that were clear and

consistently reproduced were considered. RAPD bands

were scored as present (1) or absent (0) for each DNA

sample, and a matrix of different RAPD phenotypes was

established and used for statistical analysis.

Statistical analysis of morphological traits

One-way ANOVA was performed for each of the traits,

and Duncan¡¦s test was used to test the significance of the

differences between populations. A nested ANOVA was

also performed to estimate the levels of morphological

variation within and among S. krylovii populations using

the restricted maximum likelihood method of the SAS

VARCOMP procedure (SAS, 1989). The proportion of

variance accounted for by random factor (population being

the only random factor in this analysis) was calculated

as the ratio of the variance component to the sum of all

variance components (i.e., variance among populations),

and the remainder was considered as variance within

population or variance among individuals within

population (namely error). Eucildean distance coefficients

were estimated for each pair of populations after means

of each character were normalized using Z-scores in order

to avoid effects due to scaling difference. The resulting

Eucildean distance matrix was used for cluster analysis

using the unweighted paired group method analysis

(UPGMA) (Sneath and Sokal, 1973) in NTSYS-pc (Rohlf,

1994). In order to obtain information on the traits most

effective in affecting the cluster, Principal Component

Analysis (PCA) was carried out on the mean of the

twenty morphological characters. Common components

coefficients, eigenvalues, and relative and cumulative

proportion of the total variance expressed by single traits

were calculated.

Statistical analysis of RAPD markers

The vector of each individual¡¦s RAPD marker

presence/absence was used to compute a measure of

genetic similarity for all pairs of individuals. Jaccard¡¦s

similarity coefficient (Jaccard, 1908) was used. The

Jaccard¡¦s distance was calculated by D

j

= 1- Jaccard

xy

,

and the average Jaccard¡¦s distances within and between

pairwise populations were estimated. UPGMA cluster was

generated based on the average Jaccard¡¦s distance matrix

using NTSYS-pc (Rohlf, 1994). Nei¡¦s genetic diversity

index and gene differentiation coefficient (G

ST

) were

calculated under the Hardy-Weinberg equilibrium using

the POPGENE Version 32 Program (Yeh et al., 1999).

At the same time, we used G

ST

to calculate the average

number of immigrants per generation for each locus,

namely, Nm = (1- G

ST

)/4 G

ST

.

Correlation analysis

Data used to calculate indices of climatic differences

between populations were obtained from the Xilinhot

ƒÚ

¡¦s

Meteorological Station, which was responsible for

assembling data from its subordinate meteorological

stations located in Bayanwula, Xilinhot (East), Xilinhot

(West), Xinhot, and Mandulatu, respectively. Thirty-year

average values (1965-1995) for annual precipitation¡X

.10¢XC cumulative temperature in a year, annual mean

temperature, number of frost-free days, number of windy

days, mean temperature in January and July, sunshine

hours in April and July, the percentage of sunshine time,

mean temperature 10 cm above ground level in May

and July¡Xwere compiled and transformed into standard

units (Table 3). The aridity index is defined as the ratio of

cumulative temperature in a year to total precipitation of

the year and can be viewed as an indication of the degree

of drought of an environment. It is calculated in this article

by formula: aridity index = 0.16 * cumulative temperature

in a year (.10¢XC) / precipitation in the year. The average

value, with all variables weighted equally, was used as

an index of climatic difference between sites (Knapp and

Rice, 1998). Geographic distances among populations

were estimated from the Map of the Xilingol League.

In order to examine the relationship between genetic

variations and geographical distances, and then climatic

differences among the sampling sites, the relationships

between the Jaccard¡¦s distance / Euclidean¡¦s distance and

ƒÚ

Xilinhot is the capital city of the Xilingol League, an administrative division in Inner Mongolia Autonomous Region, which

is corresponding to a prefecture.

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

27

climatic distance / geographic distance were analyzed

using a Mantel test (1967) in NTSYS-pc with significance

of the autocorrelation coefficients tested by resampling

(3000 auto permutations).

RESULTS

Morphological variations

Statistically significant differences (P<0.05) were

found among populations for the 20 morphological

traits studied (Table 4). There was no obvious trend

taking all morphological traits into account, but if we

only considered the traits related to growth (the last

ten characters), an obvious trend did emerge, and that

was that except for the length of the flag leaf blade,

plants from POP4 and POP5 were smaller than those

from other populations. This result suggested that the

characters related to growth were significantly affected

by local microhabitats and might be indicative of local

Table 3. The climatic characters in different populations of S. krylovii (mean of 30 years between 1965 and 1995).

Variables

POP1 POP2 POP3 POP4 POP5

Annual precipitation (mm)

340 300 290 230 180

.10¢XC cumulative temperature in a year (¢XC)

2256 2400 2496 2552 2664

Annual mean temperature (¢XC)

1.5 1.4 1.8 1.2 2.9

Aridity index

1.06 1.28 1.38 1.78 2.37

Number of frost free days (d)

100 106 106 103 120

Number of windy days (d)

69.9 61.1 61.1 84.6 85.5

Mean temperature in January (¢XC)

-19.4 -19.1 -19.8 -22 -18.9

Mean temperature in July (¢XC)

19.5 20.7 20.8 20.3 21.4

Sunshine hours in April (h)

261.3 266.5 269 280.4 284.6

Sunshine hours in July (h)

267.7 274.3 276.9 288.4 302.8

The percentage of sunshine time (%)

66

69

70

71

73

Mean temperature 10 cm above ground level in May (¢XC)

11 11.8 11.9 12.3

15

Mean temperature 10 cm above ground level in July (¢XC)

20 20.7

21 22.3 24.3

Table 4. Statistical analysis on the morphological characters of S. krylovii populations.

Character

POP1 POP2 POP3 POP4 POP5 Mean SE CV

Diameter of seed (mm)

0.888

a

0.829

c

0.814

c

0.668

d

0.850

b

0.810 0.084 0.104

Length of callus (mm)

2.379

c

2.496

b

2.609

a

2.215

d

2.420

bc

2.424 0.146 0.060

Length of lemma (cm)

1.075

c

1.102

b

1.145

a

1.071

c

1.062

c

1.091 0.033 0.031

Length of the first segment of awn (cm)

3.944

a

3.255

b

2.995

c

2.722

d

2.938

c

3.171 0.472 0.149

Length of the second segment of awn (cm)

1.305

c

1.400

b

1.475

a

1.377

d

1.387

a

1.389 0.061 0.044

Length of awn apex (cm)

9.928

c

10.354

b

11.370

a

9.229

d

11.660

a

10.508 1.008 0.096

Length of the first glume (cm)

2.545

c

2.578

c

2.874

a

2.300

d

2.648

b

2.589 0.206 0.080

Length of the second glume (cm)

2.402

d

2.463

c

2.697

a

2.179

e

2.521

b

2.452 0.188 0.077

Length difference of the two glumes (cm)

0.143

b

0.115

c

0.178

a

0.122

b

0.127

b

0.137 0.025 0.185

Length of awn (cm)

15.177

b

15.008

b

15.839

a

13.327

c

15.984

a

15.067 1.058 0.070

Diameter of the second internode (mm)

1.070

b

1.120

b

1.077

b

1.208

a

1.057

b

1.106 0.062 0.056

Length of the second internode (cm)

8.752

bc

10.056

a

9.346

b

8.352

c

6.584

d

8.618 1.306 0.152

Length of flag leaf blade (cm)

11.038

a

9.200

bc

9.581

bc

8.553

c

9.728 0.960 0.099

Height of reproductive shoot (cm)

47.240

a

43.980

b

46.480

a

37.580

c

34.454

d

41.947 5.655 0.135

Dry matter weight per reproductive shoot (g)

0.305

c

0.358

ab

0.398

a

0.332

bc

0.253

d

0.329 0.055 0.166

Height of vegetative shoot (cm)

23.178

a

20.820

b

24.348

a

18.896

c

16.564

d

20.761 3.155 0.152

Leaf blade length per vegetative shoot (cm) 50.466

ab

48.468

b

53.812

a

35.582

c

35.026

c

44.671 8.763 0.196

Dry matter weight per vegetative shoot (g)

0.041

b

0.043

ab

0.050

a

0.029

c

0.032

c

0.039 0.009 0.220

Length of longest leaf of vegetative shoot (cm) 21.512

a

18.260

b

22.432

a

17.340

b

15.168

c

18.942 3.002 0.159

Length of leaf sheath of first leaf (cm)

1.666

bc

2.560

a

1.916

b

1.556

c

1.396

c

1.819 0.456 0.250

28

Botanical Studies, Vol. 47, 2006

environments. Euclidean¡¦s dissimilarity / distance

coefficients were calculated for all populations based on

the morphological data in Table 4. Pairwise Euclidean

distances ranged from 0.246 to 0.633 (Table 5). Cluster

analysis placed these populations into three main

subgroups (Figure 2): POP1, POP2 and POP3 were within

the first subgroup, and there were relatively small distance

coefficients between them; POP4 and POP5 were in two

different subgroups, and both populations had relatively

large distance coefficients from any other populations.

The information on variance patterns in the characters

of S. krylovii is summarized in Table 6. The means of

coefficients of variation (CV) estimated for trait varied

from 0.058 (length of lemma) to 0.619 (length difference

of the two glumes), with the mean being 0.192. Except

for the length difference of the two glumes, the CVs

of other characters obtained from spike (the first ten

characters) were consistent between populations and

were smaller than those of characters related to growth

(the last ten characters). This result suggested that

characters related to growth had higher phenotypic

plasticity, and that characters obtained from spike were

relatively conservative and had no indication for the local

environment. The means of CV estimated for all traits

within population ranged from 0.185 (POP4) to 0.200

Table 5. Euclidean

¡¦

s distances between populations of S.

krylovii calculated from the morphological data.

Population code POP1 POP2 POP3 POP4 POP5

POP1

-

POP2

0.262 -

POP3

0.246 0.260 -

POP4

0.404 0.520 0.560 -

POP5

0.605 0.586 0.633 0.509 -

Table 6. Coefficient of variation of morphological characters in five S. krylovii populations and the variation distribution within

and among populations.

Character

Coefficient of variation

Source of variation

POP1 POP2 POP3 POP4 POP5 Mean Population Error

Diameter of seed

0.079 0.097 0.098 0.110 0.081 0.093 55.85 44.15

Length of callus

0.126 0.130 0.143 0.130 0.134 0.132 16.30 83.70

Length of lemma

0.044 0.058 0.056 0.055 0.078 0.058 20.52 79.48

Length of the first segment of awn

0.101 0.155 0.140 0.129 0.158 0.137 54.43 45.57

Length of the second segment of awn

0.105 0.136 0.110 0.121 0.146 0.124 10.06 89.94

Length of awn apex

0.125 0.130 0.122 0.098 0.092 0.113 40.87 59.14

Length of the first glume

0.077 0.090 0.089 0.064 0.072 0.078 49.39 50.62

Length of the second glume

0.073 0.089 0.091 0.072 0.068 0.079 47.49 52.52

Length difference of the two glumes

0.598 0.725 0.581 0.537 0.655 0.619 7.26 92.74

Length of awn

0.085 0.097 0.098 0.076 0.082 0.088 38.17 61.83

Diameter of the second internode

0.168 0.141 0.176 0.142 0.144 0.154 9.93 90.07

Length of the second internode

0.133 0.205 0.147 0.161 0.231 0.176 41.64 58.36

Length of flag leaf blade

0.266 0.334 0.339 0.242 0.303 0.297 8.42 91.58

Height of reproductive shoot

0.174 0.151 0.138 0.118 0.133 0.143 44.60 55.40

Dry matter weight per reproductive shoot 0.402 0.293 0.346 0.216 0.307 0.313 19.81 80.19

Height of vegetative shoot

0.138 0.167 0.165 0.162 0.158 0.158 47.05 52.95

Leaf blade length per vegetative shoot

0.165 0.208 0.200 0.231 0.180 0.197 48.84 51.16

Dry matter weight per vegetative shoot

0.271 0.345 0.365 0.458 0.271 0.342 27.40 72.60

Length of longest leaf of vegetative shoot 0.133 0.170 0.160 0.169 0.154 0.157 49.71 50.29

Length of leaf sheath of first leaf

0.448 0.284 0.394 0.404 0.426 0.391 29.16 70.84

Mean

0.186 0.200 0.198 0.185 0.194 0.192 33.34 66.66

SD

0.144 0.149 0.135 0.134 0.144 0.137 16.96 16.96

F igure 2. De ndrogram generate d with UP GMA bas ed on

Euclidean¡¦s distances of S. krylovlii populations.

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

29

(POP2), and the divergence was small; furthermore,

the difference between populations was not significant

(P>0.05). The results of the univariate analysis of variance

for the characters showed that the mean value of variation

among populations was 33.34%, ranging from 7.26% to

55.85%.

A principal components analysis of the twenty

quantitative variables yielded three statistically significant

components, which accounted for 92.08% of the total

variance (50.05% the first, 25.48% the second and 16.55%

the third). The coefficient for each variable and significant

components were shown in Table 7. The first factor was

mostly affected by length of callus, length of lemma,

length of the first and the second glume, length difference

of the two glumes, height of reproductive and vegetative

shoot, leaf blade length per vegetative shoot, dry matter

weight per vegetative shoot, and length of longest leaf of

vegetative shoot. The second factor was affected by length

of awn apex, length of awn and length of flag leaf blade.

The third factor was affected by length of the first segment

and the second segment of awn.

RAPD variations

After screening eighty 10-base oligonucleotide primers

(Operon Technologies, Inc., Alameda CA, USA) against

two bulked samples, thirteen primers that showed intense

and reproducible bands were selected for further survey.

All these primers generated 237 amplified bands in

total, and their length ranged from 300 to 2,000 bp. The

Jaccard¡¦s similarities between individual plants ranged

from 0.221 to 0.938. The average Jaccard¡¦s distances

between pairs of plants belonging to the same or to

different populations were summarized in Table 8, and a

UPGMA dendrogram based on average Jaccard¡¦s distance

Table 7. Coefficients for each morphological variable for the three significant principal components.

Variable

Component

First

Second

Third

Diameter of seed

0.547

-0.389

-0.689

Length of callus

0.914

-0.341

0.149

Length of lemma

0.869

0.133

0.466

Length of the first segment of awn

0.347

0.293

-0.869

Length of the second segment of awn

0.498

-0.277

0.813

Length of awn apex

0.482

-0.872

0.076

Length of the first glume

0.861

-0.494

0.063

Length of the second glume

0.841

-0.536

0.066

Length difference of the two glumes

0.758

-0.039

0.024

Length of awn

0.642

-0.716

-0.269

Diameter of the second internode

-0.581

0.617

0.530

Length of the second internode

0.579

0.690

0.226

Length of flag leaf blade

-0.014

0.808

-0.466

Height of reproductive shoot

0.797

0.542

-0.263

Dry matter weight per reproductive shoot

0.641

0.525

0.559

Height of vegetative shoot

0.822

0.519

-0.131

Leaf blade length per vegetative shoot

0.921

0.347

-0.161

Dry matter weight per vegetative shoot

0.985

0.133

0.018

Length of longest leaf of vegetative shoot

0.788

0.495

-0.182

Length of leaf sheath of first leaf

0.499

0.330

0.288

Percentage of Variance

50.05

25.48

16.55

Figure 3. Dendrogram generated by UPGMA based on average

Jaccard¡¦s distances of S. krylovii populations.

30

Botanical Studies, Vol. 47, 2006

between different populations was shown in Figure 3. The

average distances within the populations (boxed in Table

8) were, in all cases, smaller than the distances between

plants from different populations. This provided evidence

that individuals from different populations diverged more,

on average, than individuals within the same populations.

The range between the minimum and the maximum of the

Jaccard¡¦s genetic distances was summarized in brackets in

Table 8, and it appeared that the range within populations

was larger than that among populations.

A summary of the genetic diversity and genetic

differentiation of the five populations of S. krylovii

was given in Table 9. The Nei¡¦s genetic diversity for

individual populations (h) increased from 0.1279 (POP1)

to 0.1952 (POP5). Percentage of polymorphic bands

(PPB) increased from 41.35% (POP1) to 61.18% (POP5).

Observed number of alleles (na) increased from 1.4135

(POP1) to 1.6118 (POP5), and effective number of alleles

(ne) increased from 1.2117 (POP1) to 1.3272 (POP5). All

these parameters showed the same pattern as diversity

indexes, i.e., the parameters increased with the increase

of aridity of the study sites. Total Nei¡¦s genetic diversity

varied from 0.0112 to 0.5000, the gene differentiation

(G

ST

) based on Nei¡¦s genetic diversity (1973) was 0.3218,

and an estimate for Nm was 0.5239, indicating that 32.18%

of the genetic diversity was allocated among populations.

Correlation analysis

The relationship between two distance matrixes based

on morphological traits (Table 5) and RAPD markers

(Table 8) was estimated by Mantel¡¦s test (r = 0.3604, P

= 0.182 > 0.05, n = 3000 permutations), suggesting no

significant correlation between the two types of population

differentiation. Relationships between either of the two

distance coefficient matrixes (Euclidean¡¦s or Jaccard¡¦s

distances) of S. krylovii and geographic distance matrix

(Table 10) were tested by Mantel test, and the results were

not significant. The relationship was significant between

Euclidean¡¦s distance and climatic distance (Table 10)

(r = 0.5746, P = 0.0463, n = 3000 permutations), as well

as between Jaccard¡¦s distance and climatic distance (r =

0.7027, P = 0.0427, n = 3000 permutations) by Mantel

test, suggesting that selection associated with climatic

variation played an important role in shaping patterns of

morphological and RAPD variations in S. krylovii.

Table 8. Average Jaccard¡¦s genetic distances between pairs of individuals belonging to the same (boxed) or different populations

based on 237 RAPD markers for five populations of S. krylovii (from the minimum to maximum value of paired Jaccard¡¦s genetic

distances).

POP1

POP2

POP3

POP4

POP5

POP1 0.549(0.267-0.729)

POP2 0.812(0.700-0.923) 0.544(0.313-0.762)

POP3 0.774(0.643-0.880) 0.745(0.613-0.854) 0.531(0.221-0.700)

POP4 0.875(0.787-0.938) 0.81(0.758-0.913) 0.825(0.742-0.919) 0.567(0.356-0.705)

POP5 0.843(0.742-0.913) 0.812(0.710-0.927) 0.791(0.696-0.901) 0.78(0.714-0.874) 0.578(0.240-0.786)

Table 9. Patterns of genetic diversity of S. krylovii populations.

Population code Sample size na

ne

h

PPB (%) G

ST

Nm

POP1

18

1.4135 1.2117

0.1279

41.35

POP2

18

1.5316 1.2428

0.1505

53.16

POP3

18

1.5316 1.2646

0.1620

53.16

POP4

18

1.5527 1.2955

0.1756

55.27

POP5

18

1.6118 1.3272

0.1952

61.18

Average

1.5282 1.2684

0.1622

52.82

All populations

90

0.2392 (0.0112-0.5000) 97.47 0.3218 0.5239

Na, Observed number of alleles; ne, Effective number of alleles; h, Nei¡¦s gene diversity; PPB, Percentage of polymorphic bands; Nm,

Estimate of gene flow from G

ST.

Table 10. Geographic distances (km) (above diagonal) and

climatic factor variation coefficients (below diagonal) between

the locations of different S. krylovii populations.

POP1 POP2 POP3 POP4 POP5

POP1

- 121.90 176.60 226.54 324.47

POP2

0.0828 - 54.71 110.42 206.10

POP3

0.1110 0.0289 - 64.29 154.00

POP4

0.1950 0.1170 0.0932 - 98.13

POP5

0.2690 0.1900 0.1650 0.0772 -

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

31

DISCUSSION

Consistency of morphological and RAPD

markers

In the present study, we used not only selective markers

(morphological traits for phenotypic genetic variations)

but also selectively neutral markers (RAPD markers for

purely DNA genetic variations) to study the intra-specific

differentiation of S. krylovii and the relationships between

differentiation of S. krylovii and geographic distances and

climatic variations. With these approaches, we arrived at

the following conclusions.

First, there was significant differentiation among S.

krylovii populations both in phenotypic traits and in RAPD

markers (Tables 4, 6, 8 and 9), which consistent with

former studies on S. krylovii by Zhao et al. (2003, 2004).

Second, there was a small proportion of variance among

populations and a larger proportion of variance within

populations both in morphological traits (33.34% vs

66.656%) and in RAPD characters (32.18% vs 67.82%).

Bussell (1999) reviewed RAPD studies on population

genetics of 38 plant species and indicated that the average

genetic variance component among populations of the 30

out-breeding species was 14.4%. RAPD analysis on out-

crossing perennial grasses showed 5-15% of the genetic

variation partitioned among populations (Huff et al., 1993,

1998; Huff, 1997). Allozyme analysis of out-crossing

species tended to have 17% of the total genetic variation

residing among populations (Hamrick and Godt, 1997)

while out-crossing grasses exhibited 11% of the variation

among populations (Godt and Hamrick, 1998). Because

S. krylovii is a wind-pollinated perennial grass, about 33%

of the variation among populations is significantly higher

than other out-crossing perennial grasses in population

differentiation. The high genetic differentiation among

S. krylovii populations was also evidenced by Han et al.

(2003). To date, it is well established that the main factors

that determine the population genetic structure of plants

include the mating and reproductive system, selection or

others (Hamrick and Godt, 1989). In our case, there was

evidence that habitat destruction and degradation from

decades of over-grazing and urbanization throughout the

geographic range of S. krylovii had significantly decreased

its populations in northern China both in scale and in size

(Li, 1997) although it remained difficult to rank these

according to their significance. A rational explanation of

the present results was perhaps fragmentation of habitats

and small populations (small in size or decreased in sexual

reproduction by over-grazing) resulting from human

activities. Similar observations were reported about other

wind-pollinated grasses (Qian et al., 2001).

Third, neither phenotypic variation and geographic

distance nor molecular variation and geographic distance

showed any significant correlation, suggesting that the

differentiation of S. krylovii estimated by morphological

traits or by RAPD markers was not consistent with the

Isolation-by-distance model (Wright, 1946). However,

there was significant correlation between phenotypic

variation and climatic variation and between genetic

variation and climatic variation, which is consistent with

the hypothesis that selection associated with climatic

variation plays an important role in shaping patterns of

morphological and RAPD variations in S. krylovii. Shmida

et al. (1986) suggested that a decrease in size and organ

dimensions is a general rule for plants distributed along

a climatic gradient towards the desert. In our study, the

habitat climates of POP1 ~ POP3 were relatively humid

while POP4 and POP5 were located in the arid desert

steppe. In Table 4, the last ten characters related to growth

(except for the length difference of the two glumes) were

smaller in POP4 and POP5 than in other populations, that

is to say, the last ten characters were smaller in the desert

steppe. This might be an adaptation to aridity, presumably

for reasons of reducing water loss through reduction of

the area exposed to radiation (Shmida et al., 1986). From

a functional standpoint, it has been argued that smaller

leaves may be favored in drier environments, because

smaller leaves provide less surface area for transpirational

water loss (Givnish, 1979; Nobel, 1991; Dudley, 1996).

In addition, smaller organ size and smaller plant size can

reduce developmental time (Guerrant, 1988). Similar

phenotypic gradients could also be found in other plant

and animal groups (Endler, 1977; Nevo, 1988). Directional

change in morphological characters along climatic

gradient is natural selection, rather than random processes,

and plays a dominant role in shaping these characters

(Endler, 1977; Davis and Gilmartin, 1985).

Discrepancy of morphological and RAPD

markers

Patterns of genetic variation based on morphological

traits and RAPD markers were not consistent with each

other, and two aspects of the discrepancy should be

emphasized.

First, the pattern of the UPGMA dendrogram based on

morphological traits and the one based on RAPD markers

did not match each other. Populations from the same

habitats were clustered into one subgroup based on RAPD

markers while the clustering based on morphological

traits showed that POP1 (from the meadow steppe) and

POP3 (from the typical steppe) were clustered together

first. Moreover, the Mantel¡¦s test indicted no significant

correlation between Euclidean¡¦s distance (morphological

traits) and Jaccard¡¦s distance (RAPD markers). Some

authors found that the genetic-phenetic variations were

positively and closely correlated (Houle, 1989; Briscoe

et al., 1992; Soule and Zegers, 1996; Waldmann and

Andersson, 1998). However, widely differing opinions

about the magnitude of the relationship have been

expressed, and the extent of the correlation remains

controversial. Lewontin (1984) and Lynch (1996) pointed

out that genetic-phenotypic correlations were likely to

be low. Patterson et al. (1993) found the relationship

between molecular and morphological phylogenies to be

32

Botanical Studies, Vol. 47, 2006

weak. Reed and Frankham (2001) analyzed the correlation

coefficients of 71 datasets by a meta-analysis, and

noted that the mean of correlation coefficients between

molecular and quantitative measures of genetic variation

was weak (r = 0.217). Nevertheless, it is reasonable to

assume that the direction or magnitude of selective force

acting on the majority of RAPD variation differs from

that acting on many morphological traits. From this

discrepancy in the present study, we could infer that local

selection rather than migration or genetic drift played

a more important role in the divergence of S. krylovii

because if genetic drift was the dominant evolutionary

force leading to divergence among populations,

selectively neutral markers and selective quantitative

trait variations should have shown a stronger relationship

(Knapp and Rice, 1998). Further evidence was available

in which isozyme and morphological markers were used

to study genetic variations among populations (Bryant,

1984; Lagercrantz and Ryman, 1990). Our recent

field investigation showed that the current geographic

populations have been gradually reduced in size because

of increased disturbance from human activities. It is

well known that the measure of phenotypic variations

is environmentally dependent while molecular markers

are rather independent. Thus, phenotypic variations

were more affected than molecular variations during

the period of intensive human activities, e.g., as grazing

intensity increased, the height of plant decreased (Wang

et al., 2000). An alternative way to deal with the poor

correlation between genetic and morphological distances

was proposed by Roldan-Ruiz et al. (2001), who selected

only molecular markers linked to phenotypic traits in DUS

(Distinctness, Uniformity and Stability) testing.

Second, in the present study, both coefficients of

variation of morphological traits within populations

and Nei¡¦s genetic diversity index within populations

indicated a high diversity of S. krylovii individuals within

populations. Nei¡¦s genetic diversity increased from the

east to the west populations, but coefficients of variation

for morphological traits within populations varied without

regularity from the east to the west. A possible explanation

for this discrepancy was given in the following: i) Higher

genetic diversity within populations was associated with

higher capability in adapting to the changing conditions

(Sun, 1996; Rajora et al., 1998). In the present study, the

environmental conditions turned harsh gradually moving

from east to west. Higher genetic diversity was reserved

under adverse environmental conditions in evolutionary

history. ii) Although RAPD and morphological variations

were shaped by climatic selection, the influence of

microhabitat may be the major factor that affected

population divergence within a population. Selectively

morphological traits were environmentally dependent,

and higher CVs within populations might represent

larger microhabitat selection forces. When disturbance

increased with human activities and/or changing climates,

the selectively morphological trait variation would have

happened due to phenotypic plasticity, but variation at

DNA level would not have happened in such a short

term. This result supported a view that S. krylovii is more

responsible to local selection pressures and explained why

this species has a broad ecological amplitude.

Advantages of combination of morphological

and RAPD markers

Despite the obvious advantages in analyzing population

genetic structure that have been summarized by many

authors, disadvantages of the method (selectively neutral

markers such as RAPD and selectively quantitative traits)

have been noticed by more and more workers involved

in methodology development. RAPD markers or other

forms of neutral or nearly neutral molecular markers,

are unlikely to accurately predict patterns of variation in

quantitative traits when selection, rather than drift, is the

primary force acting (e.g., local adaptation, speciation)

(Reed and Frankham, 2001). In some cases, plant groups

with very low levels of molecular differentiation among

populations show significant levels of morphological

genetic differentiation (Furnier et al., 1991; Karhu

et al., 1996). Quantitative trait variation has several

disadvantages. Obtaining accurate data is time-consuming

and limited by growing season (Camlin and Gilliand,

1994), and the expression of quantitative traits is generally

rather plastic with respect to environmental effects. (If the

common garden experiments were selected to analyze the

quantitative trait variation in order to reduce the influence

of environmental factors, it would take more time.) Smith

and Smith (1989) suggested that the use of morphological

traits was not always the best way to evaluate genetic

distance since the degree of divergence between genotypes

at the phenotypic level is not necessarily correlated

with a similar degree of genetic difference (Hamrick

and Godt, 1989). Traditional morphological observation

alone cannot determine the roles of phenotypic plasticity

and genetic differentiation on population variation (Wen

and Hsiao, 1999). Considering the weak correlation

between molecular differentiation and quantitative genetic

variation, a combination analysis of both quantitative traits

and molecular markers might be most desirable because it

would allow examination of the relative roles of selection,

drift, and gene flow in structuring genetic variation species

(Felsenstein, 1986; Rogers, 1986; Knapp and Rice, 1998).

The combination has been used more and more in the

past several years and has proven to be powerful (Black-

Samuelsson et al., 1997; Szczepaniak et al., 2002).

In summary, the five S. krylovii populations showed

significant differentiation both in quantitative traits

and RAPD markers. It was selection, not genetic

drift, that was the major factor in this differentiation.

As for the biological characters of S. krylovii, larger

genetic differentiation suggested that genetic diversity

had been affected by the declining of population size

and fragmentation of the habitat by human activities.

Unfortunately, without historical genetic data, we do

not know how to compare the current levels of genetic

diversity with those prior to habitat reduction and

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

33

fragmentation. However, the information obtained in

the present study reminds us that it would be prudent

to work out some measures to protect S. krylovii that

would prevent it from larger differentiation in genetic

variation and in morphological variation. It is generally

believed that mutation and genetic drift due to finite

population size, and natural selection will lead to genetic

diversification of local populations and that the movement

of gamets and individuals (i.e. gene flow) will counter

that diversification. We should protect the favorite

ecological environments of S. krylovii and avoid habitat

fragmentation and degradation due to over-grazing, in

order to maximize the movement of gamet and individuals.

In conserving germplasm resources of S. krylovii, w e

should protect populations with significant differentiation

either in genetic variation or in quantitative trait variation.

In a word, the combined analysis of both morphological

and RAPD markers has undoubtedly provided important

information on S. krylovii¡¦s genetic diversity and will

provide information useful in formulating effective

conservation decisions.

Acknowledgements. This work was supported by National

Basic Research Program of China (G2000018601).

LITERATURE CITED

Baruch, Z., J.M. Nassar, and J. Bubis. 2004. Quantitative trait,

genetic, environmental, and geographical distances among

popul ations of the C

4

gras s Tr achypogon plumos us in

Neotropical savannas. Divers. Distrib. 10: 283-292.

Black-S amuels son, S ., G. Erikss on, L . Gus tafss on, and P.

Gustafsson. 1997. RAPD and morphological analysis of the

rare plant species Vicia pisiformis (Fabaceae). Biol. J. Linn.

Soc. 61: 325-343.

Bris coe, D. A., J .M. Malpica, A. Roberts on, G.J. Sm ith, R.

Frankham, R.G. Banks , and J.S.F. Barker. 1992. Rapid

loss of genetic variation in large captive populations of

Drosophila files: implications for the genetic management

of captive populations. Conserv. Biol. 6: 416-425.

Bus sel l, J.D. 1999. The distri but ion of ra ndom am plified

polymorphic DNA (RAPD) diversity amongst populations

of Isotoma petraea (Lobeliaceae). Mol. Ecol. 8: 791-802

Brya nt, E.H. 1984. A com pari son o f e le ctropho reti c and

morphometric variability in the face fly, Musca autumnlis.

Evolution 38: 455-458.

Ca ml in, M.S . and T.J. Gil lil and. 19 94. Rec ent tec hnic al

developments for pos sible use in the registration of new

plant varieties Invited paper in Biometrics in plant breeding:

application of molecular markers. In J.W. van Ooijen and

J. Jans en (eds.), Proc 9

th

Meeting EUCARPIA Se ction

Biometrics in P lant Breeding, July 1994, Wageningen,

Netherlands, pp. 8-14.

Chung, M.G., J.L. Hamirck, S.B. Jones, and G.S. Derda. 1991.

Isozyme variation within and among populations of Hosta

(Liliaceae) in Korea. Syst. Bot. 16: 667-684.

Davis, J.I. and A.J. Gilmartin. 1985. Morphological variation

and speciation. Syst. Bot. 10: 417-425.

Dudley, S.A. 1996. Differing selection on plant physiological

traits in response to environmental water availability: a test

of adaptive hypotheses. Evolution 50: 92-102.

Endler, J.A. 1977. Geographic Variation, Speciation, and Clines.

Princeton University Press, Princeton, New Jersey.

Fahima, T., G.L. Sun, A. Beharav, T. Krugman, A. Beiles, and

E. Nevo. 1999. RAPD polymorphism of wild emmer wheat

populations, Triticum dicoccoides, in Israel. Theor. Appl.

Genet. 98: 434-447.

F elsens tein, J. 1986. Population difference in quantitative

characters and gene frequencies: A comment on a paper by

Lewontin and Rogers. Am. Nat. 127: 731-732.

F ritis ch, P. and L.H. Ri esebe rg. 1996. T he use of random

amplifie d polym orphic DNA (RAP D) in cons erva tion

genetics. In T.B. Smith and R.K. Wayne (eds.), Molecular

Genetic Approaches in Conservation, Oxford University

Press, London, pp. 54-73.

Furnier, G.R., M. Stine, C.A. Mohn, and M.A. Clyde. 1991.

Geographic patterns of variation in allozymes and height

growth in white spruce. Can. J. Forest Res. 21: 702-712.

Givnish, T.J. 1979. On the adaptive significance of leaf form. In

O.T. Solbrig, S. Jain, G.B. Johnson and P.H. Raven (eds.),

Topics in Plant Population Biology. Cambridge University

Press, New York, pp. 171-214.

Godt, M.J.W. and J.L. Hamrick. 1998. Allozyme diversity in

the grasses. In G.P. Cheplick (ed.), Population Biology

of Grasses, Cambridge University Press, Cambridge, pp.

11-29.

Guerrant, E.O. 1988. Heterochrony in plants the intersection

of evolution ecology and ontogeny. In M.L. McKinney

(ed.), Heterochrony in Evolution. Plenum, New York, pp.

111-133.

Hamrick, J.L. and M.J. Godt. 1989. Allozyme diversity in plant

species. In A.D.H. Brown, M.T. Clegg, A.L. Kahler, and

B.S. Weir (eds.), Plant Population Genetics, Breeding, and

Genetic Resources, Sinauer Associated, Sunderland, MA,

pp. 43-63.

Hamrick, J.L. and M.J.W. Godt. 1997. Effects of life history

traits on genetic diversity in plant species. In J. Silvertown,

M. F ranco, and J.L. Harper (eds .), Plant Life Hisories,

Cambridge University Press, Cambridge, pp. 102-118.

Han, B., J.Wang, M.L.Zhao, Z .X. Xu, and P.F. S uo. 2003.

Genetic differentiation of Stipa krylovii in di ffe rent

degraded soil. Acta Agrestia Sin. 2: 146-153.

Hi ll is , D., C. Mori tz c , a nd B . Ma bl e. 1 996 . Mol e cul a r

Systematics. Sinauer Associated, Sunderland, MA, pp. 655.

Houle, D. 1989. Allozyme associated heterosis in Drosophila

melanogaster. Genetics 143: 1467-1483.

Huff, D.R. 1997. RAP D cha racterization of heterogeneous

perennial ryegrass cultivars. Crop Sci. 37: 557-564.

Huff, D.R., R. Peakall, and P.E. Smouse. 1993. RAPD variation

within and am ong natural populat ions of outcros sing

34

Botanical Studies, Vol. 47, 2006

buffalogram [Buchloe dactyloides (Nutt) Engelm]. Ther.

Appl. Genet. 86: 927-934.

Huff, D.R., J.A. Quinn, B. Higgins , and A.J. Palazzo. 1998.

Random amplified polymorphic DNA (RAPD) variation

among native little bluestem [Schizachyrium s coparium

(Michx.) Nash] populations from sites of high and low

fert ility in fores t and gras sland biomes . Mol. Ecol. 7:

1591-1597.

Jaccard, P. 1908. Nouvelles recherches sur la distribution florale.

Bull. Soc. Vaud. Sci. Nat. 37: 557-564.

Ka rhu , A ., P. Hu rm e, M. Ka rj al ai ne n, P. Ka rv one n, K.

Karkk aine n, S . Nea le, a nd O. S avol ain en. 199 6. Do

mol ecul ar ma rkers reflec t pat terns of differenti ati on

in adaptive traits of c onifers. Theor. Appl. Genet. 93:

215-221.

Kercher, S. and J.K. Conner. 1996. Patterns of genetic variability

within and among populations of wild radish, Raphanus

rahanistrum (Brassicaceae). Am. J. Bot. 83: 1416-1421

Knapp, E.E. and K.J. Rice. 1998. Comparis on of isozymes

and quantitative traits for evaluating patterns of genetic

variation in purple needlegrass (Nassella pulchra). Conserv.

Biol. 12: 1031-1041.

L age rcra ntz , U. and N. Ry ma n. 1990. Ge net ic s truc ture

of N or wa y s p ru c e ( P i c ea a bi e s ) : c on c or da n c e o f

morphological and allozymic variation. Evolution 44:

38-53.

L ande , R. 1988. Gene tic s a nd dem ogra phy in bi ologic al

conservation. Science 241: 1455-1460.

Le wontin, R.C. 1984. Dete cting populat ion differences in

quantitative characters as opposed to gene frequencies. Am.

Nat. 123: 115-124.

Li, B. 1997. Outline of the steppe vegetation of China and

its utilization. In Editorial of Libo¡¦s works (eds.), Libo¡¦s

Works. Sience Press, Beijing, China, pp. 399-400.

Lyn ch, M. 19 96. A qu ant it at ive -ge ne ti c p ers pe ct iv e o n

conservation issues. In J.C. Avise and J.L. Hamrick (eds.),

Conservation genetics: case histories from nature, Chapman

and Hall, New York, pp. 471-501.

Mantel, N.A. 1967. The detection of disease clustering and a

generalized regression approach. Cancer Res. 27: 209-220.

Nei, M. 1987. Molecular Evolutionary Genetics. Columbia

University Press, New York.

Nevo, E. 1988. Natural selection in action: the interface of

ecology and genetic in adaptation and speciation at the

molecular and organismal levels. In Y. Yom-Tov and E.

Tchernov (eds.), The Zoogeography of Israel, Dordrecht,

Dr. W. Junk Publishers.

Nobel, P.S . 1991. Phys iochemical and Environmental P lant

Physiology. Academic Press, San Diego.

Olfelt , J .P., G.R. F urnier, a nd J .J. L uby. 2001. Wha t data

determine whether a plant taxon is distinct enough to merit

legal prot ection. A cas e study of Sedum integrifolium

(Crassulaceae). Am. J. Bot. 88: 401-410.

P atte rs on, C., D.M. Wil liam s, and C.J. Humphries . 1993.

Co ngrue nc e b et wee n m ol ec ula r a nd m orp hol ogic a l

phylogenies. Annu. Rev. Ecol. Syst. 24: 153-188.

Qian, W., S. Ge, and D.Y. Hong. 2001. Genetic variation within

and among populations of a wild rice Oryaz granulata from

China detected by RAPD and ISSR markers. Theor. Appl.

Genet. 102: 440-449.

Rajora, O.P., L. DeVerno, A. Mosseler, and D.J. Innes. 1998.

Genetic diversity and population structure of dis junc t

Newfoundland and central Ontario populations of eastern

white pine (Pinus strobus). Can. J. Bot. 76: 500-508.

Reed, D.H. and R. Frankham. 2001. How closely correlated are

molecular and quantitative measures of genetic variation. A

meta-analysis. Evolution 56: 1095-1103.

Rogers , A.R. 1986. P opula tions diffe re nc e in quantitative

characters as opposed to gene frequencies. Am. Nat. 127:

729-730.

R ohlf, F. J. 1 994. NT S YS- pc, Num e rica l Tax onom y an d

Mul tiva ria te Anal ys is S ys te m, Ve rs ion 1 .80. Ex ete r

Software, New York.

Roldan-Ruiz, I., F.A. van Eeuwijk, T.J. Gilliland, P. Dubreuil,

C. Dillmann, J. Lallemand, M. De Loose, and C.P. Baril.

2001. A comparative study of molecular and morphological

methods of describing relations hips between perennial

ryegrass (Lolium perenne L.) varieties. Theor. Appl. Genet.

103: 1138-1150.

S AS Institute, Inc. 1989. SAS /S TAT Users Guide Version 6

Fourth Edition. Cary, North Carolina, USA.

Schonewald-Cox, C.M., S.M. Chambers, B. Macbryde, and W.L.

Thomas. 1983. Genetics and Conservation: A Reference for

Managing Wild Animal and Plant Populations. Benjamin

Cummings, Menlo Park, CA.

S hmida, A., M. Evenari, and I. Noy-Meir. 1986. Hot desert

ecosystems: an intergrated view. In M. Evenari (eds.), Hot

Deserts and Arid Shrublands, Elsevier. Science Publishers,

Amsterdam.

S m ith, J .S .C. and O.S . Sm ith. 1989. The de sc riptio n a nd

asses sment of distances between inbred lines of maize:

II. The utility of morphological, biochemical, and genetic

descriptors and a scheme for the testing of distinctiveness

between inbred lines. Maydica 34: 151-161.

S neath, P. and R. Sokal. 1973. Numerical Taxonomy. W.H.

Freeman and Company, San Francisco.

S oul e, M.E. a nd G.P. Ze gers . 1996. P hene tic s of na tural

populations. V. Genetic correlates of phenotypic variation

in the pocket gopher (T homomys bottae) in California. J.

Heredity 87: 341-350.

Sun, M. 1996. Effects of populatiom size, mating system, and

evolutionary origi n on genetic diversity in Spiranthes

sinensis and S. hongkongensis. Conserv. Biol. 10: 785-795.

S z cz ep ani ak , M., E . C ie sl ak , an d P.T. Be dna rek. 2002 .

Morphological and AFLP variation of Elymus repens (L.)

Gould (Poaceae). Cell Mol. Biol. Lett. 7: 547-558.

Wa ld ma nn , P. a nd S . An de rs s on. 199 8. C om pa ri s on o f

quantitative genetic variation and allozyme diversity within

WANG et al. ¡X Morphological and RAPD analysis of

Stipa krylolvii

35

and between populations of Scabios a canescens and S.

columbaria. Heredity 81: 79-86.

Wang, W., C.Z. Liang, Z.L. Liu, and D.Y. Hao. 2000. Analysis

of the plant individual behavior during the degradation

and re st oring s ucc es s ion in s t eppe c om muni ty. Ac ta

Phytoecologica Sin. 24: 268-274.

Wen, C.S. and J.Y. Hsiao. 1999. Genetic differentiation of Lilium

Longiflorum Thunb. var. Scabrum Masam. (Liliaceae) in

Taiwan using Random Amplified Polymorphic DNA and

morphological characters. Bot. Bull. Acad. Sin. 40: 65-71.

Williams, J.G.K., A.R. Kubelik, K.J. Livak, J.A. Rafalski, and

S.V. Tingey. 1990. DNA polym orphism s am plifie d by

arbitrary primers are useful as genetic markers. Nucl. Acids

Res. 18: 6531-6535.

Wright, S. 1946. Isolation by distance under diverse systems of

mating. Genetics 31: 39-59.

Yeh, F.C., R.C. Yang, and T. Boyle. 1999. POPGENE Version

1.32, Microsoft Window Base Software for Population

Genetic Analysis: a Quick Us er¡¦s Guide. University of

Alberta, Center for International Forestry Research, Alberta,

Canada.

Zhao, M.L., B. Han, J.F. Wu, H.L. Zhang, H. Zhen, and P.F.

Suo. 2003. A preliminary study on the population variation

of Stipa krylovii. Acta Agr. Sin. 11: 163-169.

Zhao, N.X., Y.B. Gao, J.L. Wang, L. Chen, W.B. Ruan, A.Z.

Ren, and H.F. Liu. 2004. A RAP D analysis on genetic

differentiation of Stipa kr ylov ii populations in c ent ral

and eastern Inner Mongolia Steppe. Acta Ecol. Sin. 24:

560-566.

¤º»X¥j¯ó­ìÀu¶Õ´Óª«§J¤ó°w­Tªº§ÎºA©M RAPD ¤ÀªR

¤ýª÷Às¡@°ª¥É¸¶¡@»¯©À®u¡@¥ô¦wªÛ¡@¨¿ºûÙy¡@³¯¡@½¡@¼B´º¬Â¡@§õªøªL

¤¤°ê¤Ñ¬z«n¶}¤j¾Ç¥Í©R¬ì¾Ç¾Ç°| ´Óª«¥Íª«¾Ç»P¥ÍºA¾Ç¨t

§J¤ó°w­T¡]Stipa krylovii Roshev.¡^¬O¤º»X¥j¯ó­ì¤@ºØ­«­nªº¦h¦~¥ÍÂO¥Í¥Ý¯ó¡C¥¦¤À§G­±¿n¼s¡A¨ã¦³

­«­nªº¸gÀÙ©M¥ÍºA»ù­È¡C¥»¤å¥H¨ä¤À§G°Ï¤º¤­­Ó¨å«¬¼Ë¦aªº¦ÛµMºØ¸s¬°¬ã¨s¹ï¶H¡A¤ÀªR¤F§J¤ó°w­TºØ

¸sªº¼Æ¶q©Êª¬Åܲ§©M RAPD Åܲ§¡A¨Ã§Q¥Î Mantel ÀËÅç¤ÀªR¤F§ÎºAÅܲ§»P RAPD Åܲ§¡A¥H¤Î³o¨âºØÅܲ§

¤À§O»PºØ¸s¶¡ªº¦a²z¶ZÂ÷¥H¤Î®ð­Ô¤À²§¤§¶¡ªº¬ÛÃöÃö«Y¡C¤­­ÓºØ¸sªº§ÎºA¼Æ¶q©Êª¬©M RAPD ¤À¤l¼Ð°O

µ²ªG³£ªí©ú§J¤ó°w­TºØ¸s¶¡¤À¤Æ©úÅã¡F¨âºØ¤èªk©Ò±o§J¤ó°w­TºØ¸s¶¡ªº¤À¤Æ§¡»P¦a²z¶ZÂ÷¤§¶¡¤£¦s¦b

ÅãµÛªº¬ÛÃöÃö«Y¡A¦ý»P®ð­Ô¤À²§¤§¶¡¬ÛÃöÅãµÛ¡C³o¨Çµ²ªGªí©ú§J¤ó°w­TºØ¸s¶¡ªº¤À¤Æ¤£²Å¦X "¦a²z¡Ð

¹jÂ÷¼Ò«¬"¡A¦Ó¥D­n¨ü·í¦a®ð­Ô¦]¤l¤À²§ªº¼vÅT¡C³o¤£¶È¬°¨Mµ¦ªÌ¨î©w¥_¤è¯ó­ì«OÅ@±¹¬I´£¨Ñ¦³¥Îªº«H

®§¡A¦P®É¤]¬°§J¤ó°w­T³o¤@­«­nª«ºØªººØ½¸ê·½«OÅ@´£¨Ñ°ò¦¸ê®Æ¡C

ÃöÁäµü¡G®ð­Ô¦]¤l¡F¿ò¶ÇÅܲ§¡F¦a²z¶ZÂ÷¡FMantel ÀËÅç¡F¼Æ¶q©Êª¬¡FRAPD¡F§J¤ó°w­T¡C