1322 
Genetic diversity of fringed brome (Bromus 
ciliatus) as determined by amplified fragment 
length polymorphism 
Yong-Bi Fu, Bruce E. Coulman, Yasas S.N. Ferdinandez, Jacques Cayouette, and 
Paul M. Peterson 
Abstract:   Fringed brome (Bromus ciliatus L.) is found in native stands throughout a large area of North America. Little 
is known about the genetic diversity of this species. The amplified fragment length polymorphism (AFLP) technique was 
applied to assess the genetic diversity of 16 fringed brome populations sampled in Canada from the provinces of Alberta, 
British Columbia, Quebec, and Saskatchewan. Four AFLP primer pairs were employed to screen 82 samples with four to 
six samples per population and 83 polymorphic AFLP bands scored for each sample. The frequencies of the scored bands 
in all assayed samples ranged from 0.01 to 0.99 and averaged 0.53. Analysis of molecular variance revealed that 52.6% of 
the total AFLP variation resided among the 16 populations and 20.6% among the four provinces. The five Quebec popula- 
tions appeared to be genetically the most diverse and distinct. The AFLP variability observed was significantly associated 
with the geographic origins of the fringed brome populations. These findings are useful for sampling fringed brome germ- 
plasm from natural populations for germplasm conservation and should facilitate the development of genetically diverse re- 
gional cultivars for habitat restoration and revegetation. 
Key words: native grass, fringed brome, genetic variation, AFLP, habitat restoration. 
Resume :   On retrouve le brome cilie (Bromus ciliatus L.) dans des peuplements indigenes, sur la majeure partie de 
l'Amerique du Nord. On sait peu de choses sur la diversite genetique de cette espece. Les auteurs ont utilise la technique 
du polymorphisme de la longueur des fragments d'amplification (AFLP) pour evaluer la diversite genetique de 16 popula- 
tions du brome, echantillonnees au Canada, incluant les provinces d'Alberta, de la Colombie-Britannique, de Quebec et de 
la Saskatchewan. Ils ont utilise quatre paires d'amorces, pour cribler 82 echantillons constitues de quatre a six prelevements 
par population, et il ont enregistre 83 bandes AFLP polymorphiques pour chaque echantillon. Les frequences des bandes 
obtenues chez tous les echantillons testes vont de 0,01 a 0,99, avec une moyenne de 0,53. L'analyse de la variance mole- 
culaire revele que 52,6 % de la variation AFLP totale se retrouve entre les 16 populations, et 20,6 % entre les provinces. 
Les cinq populations du Quebec semblent etre les plus genetiquement diverses. La variability AFLP observee est significa- 
tivement associee aux origines geographiques des populations du brome cilie. Ces constatations sont utiles pour echantil- 
lonner le materiel genetique provenant de populations naturelles, en vue de la conservation de ce materiel, et devraient 
rendre plus facile le developpement de cultivars genetiques regionalement diversifies, pour la restauration des habitats et de 
la vegetation. 
Mots cles : graminee indigene, brome cilie, variation genetique, AFLP, restauration des habitats. 
[Traduit par la Redaction] 
Introduction 
Native plant species have gained renewed interest in re- 
Received 9 May 2005. Published on the NRC Research Press 
Web site at http://canjbot.nrc.ca on 15 December 2005. 
Y.-B. Fu, B.E. Coulman,1 and Y.S.N. Ferdinandez. 
Agriculture and Agri-Food, Canada, Saskatoon Research Centre, 
107 Science Place, Saskatoon, SK S7N 0X2, Canada. 
J. Cayouette. National Program on Environmental Health- 
Biodiversity, Agriculture and Agri-Food, Canada, Wm Saunders 
Building (No. 49), Central Experimental Farm, Ottawa, ON K1A 
0C6, Canada. 
P.M. Peterson. Department of Botany, National Museum of 
Natural History, MRC-166, Smithsonian Institution, Washington, 
DC 20013-7012, USA. 
'Corresponding author (e-mail: bruce.coulman@usask.ca). 
search and plant breeding programs in recent years. Particu- 
lar interest includes the use of these species for soil 
stabilization (Cooper 1957), mine site reclamation (Gaffney 
and Dickerson 1987), wildlife habitat restoration (Duebbert 
et al. 1981), and development of high-quality forage (Vogel 
and Pedersen 1993). One of the major limitations to the use 
of native plant species is the lack of commercial seed quan- 
tities. Efforts to develop cultivars with improved seed pro- 
duction have increased in the United States and Canada 
(Crowle 1970; Smoliak and Johnson 1980, 1983; Jones et 
al. 1991). Many improved germplasm lines have been re- 
leased and made commercially available for rangeland resto- 
ration and large-scale revegetation (May et al. 1997; Englert 
et al. 2002). However, these improved plant materials may 
not maintain a sufficiently high level of genetic diversity 
necessary for adaptation in nonlocal environments (Knapp 
and Rice 1996; Roundy 1999; Larson et al. 2000). Such con- 
Can. J. Bot. 83: 1322-1328 (2005) doi: 10.1139/b05-112 2005 NRC Canada 
Fu et al. 1323 
cern has its roots not only in adaptation but also in genetic 
diversity, as general knowledge about genetic diversity of 
many grass species native to North America is largely lack- 
ing (Huff et al. 1998; Larson et al. 2001a; Fu et al. 2004&). 
Little attention has been paid to assess the genetic diversity 
of native grass species, particularly using molecular techni- 
ques, for the development of diverse grass germplasm with 
improved seed production. 
Fringed brome {Bromus ciliatus L.) is a native, diploid 
(2n = 14), predominantly outcrossing, tufted perennial spe- 
cies (McKone 1985; Pavlick 1995; Peterson et al. 2001). 
This species is widely distributed in North America from 
Alaska and the Northwest Territories to Newfoundland- 
Labrador, Canada, south to Maryland, North Carolina, Ten- 
nessee, Illinois, Nebraska, Colorado, and southern California 
(Wagnon 1952; Pavlick 1995; Peterson et al. 2001). It oc- 
curs in a variety of habitats such as wet meadows, stream- 
banks, bogs, thickets, and occasionally talus slopes and 
roadsides. The diversity of habitats in which B. ciliatus 
grows has stimulated some interest in the development of 
ecological cultivars that are diverse and from a relatively 
unselected seed source and in the utilization of fringed 
brome seed sources for reseeding clearcuts and disturbed 
areas (May et al. 1997; Cayouette et al. 1997). Efforts 
have been made to collect the germplasm from the species 
range in Canada (Peterson et al. 2001) and to evaluate the 
seed yield, vegetative growth, and forage quality of this 
species with the hope of generating an additional source 
of abundant, high-quality seed specifically adapted to west- 
ern Canada (Cayouette et al. 1997; May et al. 1998, 1999). 
However, little effort has been made to assess the genetic 
diversity of this species. Thus, challenges exist in improv- 
ing seed production of this species while simultaneously 
maintaining genetic diversity for adaptation in nonlocal en- 
vironments (Smith and Whalley 2002). 
The amplified fragment length polymorphism (AFLP) 
technique (Vos et al. 1995) is a robust, highly effective 
method of DNA fingerprinting that can be used to assess 
molecular genetic variability. AFLP markers, although 
scored dominantly (i.e., without distinction between homo- 
zygotes and heterozygotes) and not always homologous 
(Koopman 2005), have been successfully applied to detect 
genetic variation in many grass species, including smooth 
bromegrass {Bromus inermis Leyss.) and meadow brome - 
grass {Bromus riparius Rehmann) (Ferdinandez and Coul- 
man 2002), bluebunch wheatgrass {Pseudoroegneria spicata 
[Pursh] A. Love) (Larson et al. 2000), bluegrasses {Poa 
spp.) (Larson et al. 200lb), crested wheatgrass {Agropyron 
spp.) (Mellish et al. 2002), blue grama {Bouteloua gracilis 
[Kunth] Lag. ex Griffiths) (Fu et al. 2004a), and little blue- 
stem {Schizachyrium scoparium [Michx.] Nash) (Fu et al. 
2004&). Some studies also assessed the correlation of AFLP 
variability with geographic origin of natural populations 
(Larson et al. 2001/?; Larson et al. 2003; Fu et al. 2004&). 
In a companion study aimed at clarifying B. ciliatus and 
Bromus richardsonii Link as distinct species, Peterson et al. 
(2001) revealed large AFLP variability in fringed brome, but 
no diversity analysis was specifically made for fringed 
brome populations. 
The objective of this study was to assess the patterns of 
genetic variability in fringed brome populations using AFLP 
markers with the hope to facilitate the development of di- 
verse fringed brome germplasm for revegetating clearcuts 
and other disturbed areas in Canada. 
Materials and methods 
Plant materials 
Seeds of fringed brome plants were collected at 16 loca- 
tions across Canada (five locations in Alberta, three in Brit- 
ish Columbia, five in Quebec, and three in Saskatchewan; 
see Table 1). Most of the collections were made between 
1993 and 1999. On average, seeds were collected from 10 
plants that were at least 5 m apart at each location. Col- 
lected seeds were germinated and four to six seedlings for 
each location were randomly selected. Such a small sample 
size may introduce some bias in the inference of within-pop- 
ulation variability, but population and regional comparisons 
of genetic variability should be relatively robust. Young leaf 
tissue was individually collected, freeze-dried, and stored at 
-80 ?C before DNA analysis. 
DNA extraction and AFLP procedure 
Two leaves of each sample were placed in a 2 mL 
microcentrifuge tube with two 2 mm glass beads. The 
tubes were placed on a horizontal shaker until the leaf tis- 
sue was ground to a fine powder. The DNA was extracted 
by using a DNeasy? Plant Mini Kit (QIAGEN Inc, Mis- 
sissauga, Ontario) according to the manufacturer's direc- 
tions. Total genomic DNA was quantified by fluorimetry 
using Hoechst 33258 stain (Sigma Chemical Co., St. Louis, 
Missouri). 
The AFLP analysis was performed as described by Vos et 
al. (1995) using the AFLP? Analysis System 1 provided by 
Life Technologies (Burlington, Ontario). Initially, a restric- 
tion digestion of 250 ng of genomic DNA with EcoRI and 
Msel restriction enzymes were carried out followed by liga- 
tion of adapters to the restriction fragments. This was fol- 
lowed by preamplification of the primary templates with 
AFLP primers with an additional single nucleotide at the 3' 
end. A selective amplification of the preamplified fragments 
with labeled [y-33P]EcoRI primers having three selective nu- 
cleotides at the 3' end and Msel primers having three selec- 
tive nucleotides at the 3' end was performed. The selective 
amplification was performed in an MJ Research PTC-200 
DNA engine thermocycler using the following amplification 
profile: one cycle with a denaturation step at 94 ?C for 30 s, 
an annealing step at 65 ?C for 30 s, and an extension step at 
72 ?C for 60 s. The next 12 cycles emulated a Touchdown 
polymerase chain reaction format decreasing the annealing 
temperature by 0.7 ?C each cycle to 56 ?C. Twenty-three 
more cycles were performed using the profile 94 ?C for 30 
s, 56 ?C for 30 s, and 72 ?C for 60 s. Finally, the amplifica- 
tion products were separated on 5% poly aery lamide gels for 
2.30 h at 90 W. The gels were transferred to Whatman paper 
and dried on a gel dryer for 2 h at 80 ?C. Gels were exposed 
to Kodak BIOMAX film at 80 ?C for 1-7 d depending on 
the signal intensity. From the previous AFLP analysis of 
Bromus plants (Ferdinandez and Coulman 2002), four most 
informative EcoRI-Msel primer combinations were selected 
and applied in this study (Table 2). 
2005 NRC Canada 
1324 Can. J. Bot. Vol. 83, 2005 
Table 1. Sampling site and variation pattern of AFLP for the 16 fringed brome populations studied. 
Location (code) Latitude Longitude Elevation (m)     NP      NPB      MBF       WPV 
St-Denis, Quebec (DEN) 
Levis, Quebec (LEV) 
ND-des-Bois, Mt Marbre, Quebec (MAR) 
Pontiac-Station, Quebec (PON) 
Riviere Ste-Anne-des-Monts, Quebec (RSA) 
Brancepeth, Saskatchewan (BRA) 
Okla, Saskatchewan (OKL) 
Willowbrook, Saskatchewan (WIL) 
Bezanson, Alberta (BEZ) 
Hondo, Alberta (HON) 
Nestow, Alberta (NES) 
Athabaska River, Smith, Alberta (SMI) 
Viking, Alberta (VIK) 
Chetwind, British Columbia (CHE) 
Dome Creek, British Columbia (DOM) 
Shelley, British Columbia (SHE) 
47?23'00"N 69?55'00"W 20 
46?40'00"N 71?11'00"W 100 
45?19'50"N 71?01'15"W 650 
45?28'00"N 76?18'20"W 300 
49?03'50"N 66?29'25"W 200 
53?02'10"N 105?16'10"W 440 
51?59'10"N 103?06'05"W 617 
51?11'30"N 102?48'10"W 540 
55?13'52"N 118?17'10"W 587 
55?03'15"N 114?03'20"W 610 
54?14'25"N 113?34'45"W 628 
55?04'25"N 114?05'45"W 564 
53?02'40"N 113?30'12"W 701 
55?45'10"N 121?36'10"W 914 
53?42'25"N 121?03'00"W 701 
53?56'30"N 122?35'25"W 690 
5 29 0.545 14.20 
4 6 0.500 3.33 
4 8 0.375 4.33 
4 11 0.576 5.67 
6 29 0.494 12.00 
6 18 0.454 7.53 
6 28 0.512 12.93 
4 26 0.510 14.67 
6 14 0.643 6.40 
5 16 0.475 7.40 
6 23 0.478 10.33 
5 27 0.615 12.40 
4 15 0.633 8.33 
5 25 0.520 13.00 
6 23 0.457 8.87 
6 20 0.408 8.20 
Note: NP, the number of plants assayed; NPB, the number of polymorphic AFLP bands scored; MBF, mean band frequency; WPV, within- 
population variation calculated from the sum of squares of AMOVA. The voucher information on these populations is described in Peterson et al. 
(2001) with the same code (or geographical identifier) as used in this study. 
Data analysis 
For gels generated from each primer pair, the numbers of 
observable and monomorphic AFLP bands were counted. 
Polymorphic AFLP bands with sufficient intensity for all of 
the samples were selected and manually scored as present 
(1) or absent (0). The selection may introduce bias into 
some of the AFLP variability measurements, but these se- 
lected bands should provide a relative measure of polymor- 
phism for group comparison. The selected polymorphic 
bands were analyzed for each primer pair for the level of 
polymorphism by counting the number of polymorphic 
bands and generating the summary statistics on the band fre- 
quencies. To visualize the variation pattern, the numbers of 
polymorphic bands were plotted against their frequencies of 
occurrence in all samples. For each population, the number 
of polymorphic bands, the mean band frequency, and 
within-population variation measured from the sum of 
squares from the analysis of molecular variance (AMOVA) 
(Excoffier et al. 1992) were calculated and regressed over 
the population latitude, longitude, elevation, and sample 
size using the SAS regression program (PROC REG) (SAS 
Institute Inc. 2004). 
To assess AFLP variations across various groups (popula- 
tion and region), an AMOVA was performed using Arlequin 
version 2.001 (Schneider et al. 2002). This analysis not only 
allows the partition of the total AFLP variation into within- 
and among- group variation components but also provides a 
measure of intergroup genetic distances as the proportion of 
the total AFLP variation residing between any two groups 
(called Phi statistic; Excoffier et al. 1992; Huff et al. 1998). 
Models involving various types of structuring (population 
and region) were applied. Significance of resulting variance 
components and intergroup genetic distances was tested with 
10000 random permutations. In this study, populations were 
grouped by province to form four regions for comparing re- 
gional genetic variability. 
To assess the genetic associations of the fringed brome 
plants  representing   16  populations  and four regions,  the 
bootstrapped unweighted pair group with arithmetic aver- 
ages (UPGMA) dengrograms were generated based on the 
pairwise Fst values that were obtained using the method of 
Reynolds et al. (1983) invoked in PHYLIP version 3.64 
(Felsenstein 2005) from band frequencies. The association 
of the distance matrix of the Phi statistic for 16 populations 
with the corresponding geographical distance matrix was 
also examined with the MXCOMP program of NTSYS-pc 
2.01 (Rohlf 1997) with 10000 random permutations. To 
evaluate the genetic associations of 82 individual plants, an 
individual pairwise similarity matrix was generated using 
simple matching coefficient (Sokal and Michener 1958) and 
converted to the Euclidean distance matrix for a principal 
coordinate analysis using the NTSYS-pc program. The first 
three resulting principal coordinate scores were plotted to 
assess the clustering of individual fringed bromes. 
Results 
Four AFLP primer pairs amplified a total of 327 observ- 
able bands, with the number of observable bands per primer 
pair ranging from 47 to 104 (Table 2). There were 206 
(63%) observable bands that were polymorphic, indicating 
the presence of large AFLP variation in this native grass 
species. Since many of the amplified polymorphic bands 
lacked clarity, only 83 of them were scored for further anal- 
yses. These scored bands were presumably sampled from the 
whole fringed brome genome, but their exact genome cover- 
age is unknown. The frequencies of the scored bands in all 
assayed samples ranged from 0.01 to 0.99 and averaged 
0.53. A large proportion (61%) of the scored bands had fre- 
quencies of either less than 0.16 or greater than 0.84. For 
each primer pair, statistics (mean, minimum, and maximum) 
of the band frequencies are given in Table 2. The lowest 
mean frequency was found with the primer pair E+ACG/ 
M+CTG, and the highest mean frequency was observed 
with the primer pair E+AAG/M+CAC. 
Genetic variations for 16 fringed brome populations were 
quantified in this study by the number of scored polymorphic 
2005 NRC Canada 
Fu et al. 1325 
Table 2. AFLP in 82 fringed brome plants as revealed by four AFLP primer 
pairs. 
No. of AFLP bands Frequency of scored bands 
Primer pair Total Mono Scored Mean Min. Max. 
E+ACG/M+CTG 47 14 25 0.485 0.012 0.988 
E+AAG/M+CAC 104 33 14 0.615 0.024 0.976 
E+AGG/M+CGC 87 25 14 0.512 0.012 0.988 
E+AAC/M+CAG 89 24 30 0.508 0.012 0.988 
All 327 121 83 0.530 0.012 0.988 
Note: Four EcoRl primers (E plus three selective nucleotides, 5-GACTGCGTAC- 
CAATTC+ACG, AAG, AGG, and AAC) and four Msel primers (M plus three selective 
nucleotides, 5'-GATGAGTCCTGAGTAA+CTG, CAC, CGC, and CAG) are shown for 
the primer pair combinations applied in this study. Total, total number of AFLP bands 
observed; mono, number of monomorphic AFLP bands detected; scored, number of 
polymorphic AFLP bands scored. 
bands, the mean band frequency, and the within-population 
variation measured from the AMOVA sum of squares 
(Table 1). Within a population, the number of scored poly- 
morphic bands ranged from 6 to 29 with an average of 19.9, 
the mean band frequency ranged from 0.38 to 0.64 with an 
average of 0.51, and the within-population variation ranged 
from 3.33 to 14.67 with an average of 9.35. The Quebec 
populations had the most and the least polymorphic bands 
(29 for St-Denis and Riviere de Ste-Anne-des-Monts, six for 
Levis, and eight for ND-des-Bois). The Bezanson population 
in Alberta had the highest mean band frequency, and the 
ND-des-Bois population from Quebec had the lowest mean 
band frequency. The population with the most within-popu- 
lation variation was Willowbrook (14.7), Saskatchewan, fol- 
lowed by St-Denis (14.2), Quebec. The population with the 
least within-population variation was Levis (3.3) followed 
by ND-des-Bois (4.3), Quebec. Linear regression analyses 
revealed nonsignificant associations of population latitude, 
longitude, elevation, and sample size with the population es- 
timates of the three genetic parameters mentioned (results 
not shown). 
Partitioning of the total AFLP variation into within- and 
among- population components by AMOVA showed that 
47.4% of the total variation was present within populations 
and 52.6% resided among 16 populations (Table 3). These 
variation components were significantly different from zero 
(P < 0.0001) based on the permutation test. The largest be- 
tween-population difference measured by interpopulation 
distance was observed between two Quebec populations 
(Levis and ND-des-Bois; 0.83), and the least between-popu- 
lation difference measured was between the Okla population 
in Saskatchewan and the Athabaska River population in Al- 
berta (0.09). These differences can also be visualized in the 
inferred genetic relationships of the 16 populations given in 
a dendrogram (Fig. 1A). Two populations from Quebec 
(Levis and ND-des-Bois) represented the first two most dis- 
tinct clusters with one member each. The third cluster con- 
sisted of the other three populations from Quebec, and the 
fourth cluster included 11 populations from the other three 
provinces. This pattern of clustering was similar to those 
made based on Euclidean distance (shown in Fig. 17 of Pe- 
terson et al. 2001). The AFLP variations of pairwise popula- 
tions were significantly associated (r = 0.68, P < 0.0001) 
with geographical origin of the populations. Figure 2 illus- 
trates the associations between genetic distances measured 
by the Phi statistic and geographic distances in kilometres 
of fringed brome populations. Further partitioning of the 
among-population variation into among-region and among- 
population, within-region components revealed that 20.5% 
of the total variance resided among four geographic regions 
(Alberta, British Columbia, Quebec, and Saskatchewan). 
Further assessment of the variation among the four regions 
showed that the fringed plants from Quebec were genetically 
the most distant (or distinct) from the remaining plants as- 
sayed followed by those from British Columbia (Fig. IB). 
To understand the genetic structure observed in fringed 
brome, the genetic associations of individual plants were as- 
sessed. The plot of the first two principal component scores 
based on the Euclidean distances converted from the simple 
matching coefficient matrix of 83 AFLP bands revealed 
more or less three major groups (Fig. 3). The first group on 
the upper left of the figure consisted of 58 fringed brome 
plants from Alberta, British Columbia, and Saskatchewan. 
The second group on the upper right of the figure consisted 
of 15 plants from the three Quebec populations with higher 
latitudes (Riviere Ste-Anne-des-Monts, St-Denis, and Levis). 
The third group on the lower part of the figure included one 
plant collected from Willowbrook, Saskatchewan, and eight 
plants from the two Quebec populations with lower latitudes 
(ND-des-Bois and Pontiac-Station). Clearly, the Quebec 
fringed bromes were not only diverse but also distinct from 
those fringed bromes collected from western Canada. 
Discussion 
This study represents the first attempt using AFLP 
markers to characterize the genetic variability and structure 
of Canadian fringed brome populations with the goal to en- 
hance the development of fringed brome germplasm for hab- 
itat restoration. The study revealed several interesting 
patterns. First, 52.6% of the AFLP variation was present 
among the 16 populations examined, and 20.6% resided 
among the four provinces. Second, the five Quebec popula- 
tions were both the most diverse and distinct. Third, the be- 
tween-population AFLP variability was significantly 
associated with the geographic origins of the fringed brome 
populations. These findings are useful for sampling fringed 
brome germplasm from natural populations for germplasm 
conservation and should facilitate the improvement of 
fringed brome germplasm for habitat restoration. 
2005 NRC Canada 
1326 Can. J. Bot. Vol. 83, 2005 
Table 3. AMOVA based on AFLP data among 82 fringed brome plants with re- 
spect to region and population. 
Source of variation 
Among regions 
Among populations within regions 
Within populations 
Variance 
df SS components % variation 
3 190.87 2.06? 20.52 
12 253.00 3.22a 32.10 
66 314.17 4.76? 47.38 
Note: Populations were grouped by province to form a region. 
"Significant at P < 0.0001, as calculated from 10000 random permutations. 
Fig. 1. UPGMA dendrograms illustrating the genetic relationships 
of fringed brome plants representing (A) 16 populations and (B) 
four geographic regions (BC, British Columbia; QC, Quebec; AB, 
Alberta; SK, Saskatchewan). Population labels are given in Table 1. 
Pairwise Fst genetic distances were calculated following Reynolds 
et al. (1983). Bootstrap values higher than 50 (out of 100 repli- 
cates) are indicated inside the node. 
(A) 16 populations 
LEV-QC 
  MAR-QC 
 PON-QC 
  DEN-QC 
 RSA-QC 
  DOM-BC 
 HON-AB 
100 
100 
63 
54 
93 
55_ 
0.1 
(B) four regions 
100 
100 
97 
SKE-BC 
BEZ-AB 
VIK-AB 
NES-AB 
CHE-BC 
WIL-SK 
BRA-SK 
OKL-SK 
SMI-AB 
QC 
0.1 
Fringed brome is known to be an outcrossing grass spe- 
cies also with self-compatibility (McKone 1985), although 
the level of self-fertilization in this species is not known. 
Thus, it is not fully surprising to observe the low within-pop- 
ulation variation (47.4%) shown in these populations (Ham- 
rick and Godt 1989), but a question remains on how much 
such a low estimate was accounted for by self-pollination. 
Clearly, a comprehensive study on the mating system of 
Fig. 2. Associations between genetic distances measured by the Phi 
statistic and geographic distances in kilometres of fringed brome 
populations. A linear regression is also presented. 
? 
0.8 - ? ?        ? 
?S o-7 - ? ? 
? ? 
->?? 
ts \>>^ 
.?   0.6 - * ? ^<\      ?? 
fe ? m     ^^ -*^    ? *??           ? 
?   05- ? ? ? f ? c 
C3 f /? ? -^^^        i 
.S3 ? ? ? 
"?   0.4 - ? ; ? 
u a 
? #     ? 
? 
? .? 
$  0.3- ? ? 
? 
? ? ? 
? r=0.00009X+0.31762 
tf2=0.4561, P<0.0001 
0.2 - 
? .? 
? 
0.1 - ? i i                 | 
0 500 1000 1500 2000 2500 3000 3500 4000 
Geographic distance (km) 
Fig. 3. Plot of the first two principal component scores based on 
the Euclidean distances converted from the simple matching coeffi- 
cient matrix of 83 AFLP bands for 82 fringed brome plants. These 
two components accounted for 38.6% and 14.6% of the total AFLP 
variance, respectively. Individual plants were separately labeled for 
their geographic regions (QC, Quebec; SK, Saskatchewan; AB, Al- 
berta; BC, British Columbia). 
o 
u 
0.15 
V 
V 
0.10 V 
V 
A 
'AA 
A        \A 
?         ? ^ 
0.05 ? 
0.00 ? 
0.05 ? 
V 
A 
A 
A 
A 
V        V 
?  
v
 >? 
O 
O 
? \ 
^.      ^ 
?    QC 
O    SK 
A    AB 
V    BC 
0.10 ? A ? ? 
? 
0.15 ? O 
?? 
-0.1 0.0 0.1 
PCOl (38.6%) 
0.2 
this species is needed for understanding this pattern of var- 
iation (Barrett et al. 2004). Also, it is possible that this low 
estimate may reflect the bias from the use of small sam- 
ples for these populations. Larger sample size (>30 plants 
2005 NRC Canada 
Fu et al. 1327 
per population) is needed for a more informative inference 
of within-population variation. 
The extensive differentiations observed over the popula- 
tions and regions may reflect limited gene flow via pollen 
and seed among the populations and (or) strong differential 
selection pressures existing among the sampled sites. Further 
studies are needed to provide information on gene flow be- 
tween natural fringed brome populations and its effect on 
genetic diversity (Hutchison and Templeton 1999). Also, 
the high among-population or low within-population varia- 
tions observed may not be directly associated with adaptive 
variations that exist in these populations, as AFLP markers 
presumably were selectively neutral (Vos et al. 1995). 
The within-population AFLP variability was not associ- 
ated with the population latitude, longitude, and elevation, 
probably owing to the use of small sample sizes for these 
populations. The significant association of the between-pop- 
ulation AFLP variation with geographic origin of the popu- 
lations surveyed may reflect that local adaptations existed in 
these populations, particularly in those of northern Quebec. 
A similar phylogeographic pattern was observed in Nassella 
pulchra (Hitchc.) Barkworth (Larson et al. 200lb), Elymus 
spp. (Larson et al. 2003), and little bluestem (Fu et al. 
2004&). However, the generality of this association remains 
to be determined, as the populations assessed in this study 
were located in the northern fringe of the species' range. 
Also, further assessment in a breeding nursery on adaptive 
genetic variations in reproductive growth and seed yield is 
still needed for these fringed brome populations. Such adap- 
tive assessment, currently underway, could not only help in 
the understanding of the nature of adaptation in fringed 
brome but also would facilitate the selection of fringed 
brome germplasm for habitat restoration. 
The patterns of AFLP variation observed in this study are 
comparable only with some reports for other native plants. 
For example, 57% AFLP variation was observed among 58 
natural populations of Echinacea (Mechanda et al. 2004) 
and 48% among six populations of arctic bramble {Rubus 
arcticus L. subsp. arcticus) (Lindqvist-Kreuze et al. 2003). 
In contrast, 19% AFLP variation was found among three 
populations of Festuca campestris Rydb. (Fu et al. 2005) 
and 7% among six populations of little bluestem (Fu et al. 
2004&). It is difficult to explain such discrepancy only from 
modes of reproduction (Larson et al. 2001a, b) and also is 
beyond the scope of this study. However, it is certain that 
the applications of the AFLP technique can provide useful 
measures of genetic variability in native plants, even with 
the issues of dominance and homology (Koopman 2005). 
The results of this study have implications for sampling 
fringed brome germplasm from natural populations for 
germplasm conservation. The low variation detected within 
single natural populations suggests that a rather large num- 
ber of plants from single sites should be sampled to capture 
substantial genetic variation, but the optimal number of 
plants to be sampled from one site remains to be determined 
empirically. The observation of the large between-population 
variation implies a low genetic similarity between nearby 
native stands, and thus, more distant populations should be 
sampled to capture more localized germplasm, which is 
important for the development of diverse, well-adapted 
fringed brome germplasm for habitat restoration. The find- 
ing of the large regional variation (20.6%) underscores the 
need for germplasm collection across the species' range, as 
the current sampling was applied only to the populations in 
the northern part of the range and fringed brome plants in 
more southern latitudes may also harbour substantial ge- 
netic diversity. 
The findings presented here appear to suggest the impor- 
tance of developing localized germplasm that could capture 
the unique adaptations to a particular geographic area. Thus, 
ecological varieties specific to eastern and western Canada 
should be developed for habitat restoration. Emphasis should 
be placed on the balance of local adaptation and population 
differentiation by including more germplasm from distant 
populations to sample more among-population variability. 
To capture more genetic diversity in an ecological variety, 
equal representation of nearby seed sources could be applied 
when information is limited on adaptation and diversity (Fu 
et al. 2004a). To maximize the genetic diversity in a devel- 
oped germplasm, proportional contributions of different seed 
sources based on their adaptability and genetic variability 
should be considered. As demonstrated here, the application 
of a molecular technique cannot only assess the genetic var- 
iability within and among fringed brome populations but 
also facilitate the selection of a fringed brome seed source 
that was either genetically diverse or distinct for the devel- 
opment of fringed brome germplasm for habitat restoration 
in Canada. 
References 
Barrett, S.C.H., Cole, W.W., and Herrera, CM. 2004. Mating pat- 
terns and genetic diversity in the wild daffodil Narcissus longis- 
pathus (Amaryllidaceae). Heredity, 92: 459^165. doi: 10.1038/ 
sj.hdy.6800441. 
Cayouette, J., Fillion, N., Coulman, B., and Michaud, R. 1997. 
New opportunities for native perennial bromegrasses. In Sum- 
mary of 1997 Western Forage-Beef Network Meeting. Edited 
by S.B. Acharya. Lethbridge Research Centre, Agriculture and 
Agric.-Food Canada, Lethbridge, Alta. pp. 4-5. 
Cooper, H.W. 1957. Some plant materials and improved techniques 
used in soil and water conservation in the Great Plains. J. Soil 
Water Conserv. 12: 163-168. 
Crowle, W.L. 1970. Revenue slender wheatgrass. Can. J. Plant Sci. 
50: 748-749. 
Duebbert, H.F., Jacobson, E.T., Higgins, K.F., and Podoll, 
E.B. 1981. Establishment of seeded grasslands for wildlife habi- 
tat in the prairie pothole region. US Fish Wildl. Serv. Wildl. 
Leafl. No. 234. 
Englert, J.M., Kujawski, J.L., and Scheetz, J.G. 2002. Improved 
plant materials released by NRCS and cooperators through Sep- 
tember 2001. USDA-NRCS National Plant Materials Center, 
Beltsville, Md. 
Excoffier, L., Smouse, P.E., and Quattro, J.M. 1992. Analysis of 
molecular variance inferred from metric distances among DNA 
haplotypes: application to human mitochondrial DNA restriction 
data. Genetics, 131: 479^191. 
Felsenstein, J. 2005. PHYLIP version 3.64. Available from http:// 
evolution.genetics.washington.edu/phylip.html. 
Ferdinandez, Y.S.N., and Coulman, B.E. 2002. Evaluating genetic 
variation and relationships among two bromegrass species and 
their hybrid using RAPD and AFLP markers. Euphytica, 125: 
281-291.   doi: 10.1023/A: 1015857614809. 
Fu, Y.B., Ferdinandez, Y.S.N., Phan, A.T., Coulman, B., and Ri- 
2005 NRC Canada 
1328 Can. J. Bot. Vol. 83, 2005 
chards, K.W. 2004a. AFLP variation in four blue grama seed 
sources. Crop Sci. 44: 283-288. 
Fu, Y.B., Phan, A.T., Coulman, B., and Richards, K.W. 2004i. Ge- 
netic diversity in natural populations and corresponding seed 
collections of little bluestem as revealed by AFLP markers. 
Crop Sci. 44: 2254-2260. 
Fu, Y.B., Thompson, D., Willms, W., and Mackay, M. 2005. Long- 
term grazing effects on genetic variability in mountain rough 
fescue. Rangeland Ecol. Manage., In press. 
Gaffney, F.B., and Dickerson, J.A. 1987. Species selection for re- 
vegetating sand and gravel mines in the Northeast. J. Soil Water 
Conserv. 42: 358-361. 
Hamrick, J.L., and Godt, M.J. 1989. Allozyme diversity in plant 
species. In Plant population genetics, breeding and genetic re- 
sources. Edited by A.H.D. Brown, M.T. Clegg, A.L. Kahler, and 
B.S. Weir. Sinauer Associates, Sunderland, Mass. pp. 43-63. 
Huff, D.R., Quinn, J.A., Higgins, B., and Palazzo, A.J. 1998. Ran- 
dom amplified polymorphic DNA (RAPD) variation among na- 
tive little bluestem [Schizachyrium scoparium (Michx.) Nash] 
populations from sites of high and low fertility in forest and 
grassland biomes. Mol. Ecol. 7: 1591-1598. doi: 10.1046/j. 
1365-294x.l998.00473.x. 
Hutchison, D.W., and Templeton, A.R. 1999. Correlation of pair- 
wise genetic and geographic distance measures: inferring the re- 
lative influences of gene flow and drift on the distribution of 
genetic variability. Evolution, 53: 1898-1914. 
Jones, T.A., Neilson, D.C., and Carlson, J.R. 1991. Development of 
a grazing-tolerant native grass for revegetating bluebunch 
wheatgrass sites. Rangelands, 13: 147-150. 
Knapp, E.E., and Rice, K.J. 1996. Genetic structure and gene flow 
in Elymus glaucus (blue wildrye): implications for native grass- 
land retsoration. Restor. Ecol. 4: 1-10. 
Koopman, W.J.M. 2005. Phylogenetic signal in AFLP data sets. 
Syst. Biol. 54: 197-217.   doi: 10.1080/10635150590924181. 
Larson, S.R., Jones, T.A., Hu, Z.-M., McCracken, C.L., and Pa- 
lazzo, A. 2000. Genetic diversity of bluebunch wheatgrass culti- 
vars and a multiple-origin polycross. Crop Sci. 40: 1142-1147. 
Larson, S.R., Cartier, E., McCracken, C.L., and Dyer, D. 2001a. 
Mode of reproduction and amplified fragment length poly- 
morphism variation in purple needlegrass (Nassella pulchra): 
utilization of natural germplasm sources. Mol. Ecol. 10: 1165- 
1177.   doi: 10.1046/j. 1365-294X.2001.01267.x. 
Larson, S.R., Waldron, B.L., Monsen, S.B., St. John, L., Palazzo, 
A.J., McCracken, C.L., and Harrison, R.D. 20016. AFLP varia- 
tion in agamospermous and dioecious bluegrasses of western 
North America. Crop Sci. 41: 1300-1305. 
Larson, S.R., Jones, T.A., McCracken, C.L., and Jensen, K.B. 
2003. Amplified fragment length polymorphism in Elymus ely- 
moides, Elymus multisetus, and other Elymus taxa. Can. J. Bot. 
81:789-804.   doi: 10.1139/b03-077. 
Lindqvist-Kreuze, H., Koponen, H., and Valkonen, J.P.T. 2003. 
Genetic diversity of arctic bramble (Rusbus arcticus L. subsp. 
arcticus) as measured by amplified fragment length polymorph- 
ism. Can. J. Bot. 81: 805-813.   doi: 10.1139/b03-072. 
May, K.W., Wark, B., and Coulman, B. 1997. Ecovar development 
on the northern great plains of North America. In Proceedings of 
the XVII International Grasslands Congress, Winnipeg, Man., 
and Saskatoon, Sask. Section 1.89-1.90. Association Manage- 
ment Centre, Calgary, Alta. 
May, K.W., Stout, D.G., Willms, W.D., Mir, Z., Coulman, B., 
Fairey, N.A., and Hall, J.W. 1998. Growth and forage quality of 
three Bromus species native to western Canada. Can. J. Plant 
Sci. 78: 597-603. 
May, K.W., Willms, W.D., Stout, D.G, Coulman, B., Fairey, N.A., 
and Hall, J.W. 1999. Seed yield of three Bromus species native 
to western Canada. Can. J. Plant Sci. 79: 551-555. 
Mechanda, S.M., Baum, B.R., Johnson, D.A., and Arnason, J.T. 
2004. Analysis of diversity of natural populations and commer- 
cial lines of Echinacea using AFLP. Can. J. Bot. 82: 461^-84. 
doi: 10.1139/b04-006. 
McKone, M.J. 1985. Reproductive biology of several bromegrasses 
(Bromus): breeding systems, pattern of fruit maturation, and 
seed set. Am. J. Bot. 72: 1334-1339. 
Mellish, A., Coulman, B., and Ferdinandez, Y. 2002. Genetic rela- 
tionships among selected crested wheatgrass cultivars and spe- 
cies determined on the basis of AFLP markers. Crop Sci. 42: 
1662-1668. 
Pavlick, L.E. 1995. Bromus L. of North America. Royal British 
Columbia Museum, Victoria, B.C. 
Peterson, P.M., Cayouette, J., Ferdinandez, Y.S.N., Coulman, B.E., 
and Chapman, R.E. 2001. Recognition of Bromus richardsonii 
and B. ciliatus: evidence from morphology, cytology, and DNA 
fingerprinting (Poaceae: Bromeae). Aliso, 20: 21-36. 
Reynolds, J., Weir, B., and Cockerham, C. 1983. Estimation of the 
coancestry coefficient: basis for a short-term genetic distance. 
Genetics, 105: 767-779. 
Rohlf, F.J. 1997. NTSYS-pc 2.1. Numerical taxonomy and multi- 
variate analysis system. Exeter Software, New York. 
Roundy, B.A. 1999. Lessons from historical rangeland revegetation 
for today's restoration. In Revegetation with native species: 
Rocky Mountain Research Station Proceedings No. 8. Edited by 
L.K. Holzworth, and RW. Ray. USDA Forest Service, Rocky 
Mountain Research Station, Ogden, Utah. pp. 33-38. 
SAS Institute Inc Inc 2004. The SAS system for windows V8.02. 
SAS Institute Inc., Cary, N.C. 
Schneider, S., Roessli, D., and Excoffier, L. 2002. Arlequin ver 
2.001: a software for population genetics data analysis. Genetics 
and Biometry Laboratory, University of Geneva, Geneva, Swit- 
zerland. Available from http://lgb.unige.ch/arlequin/software/ 
(verified 4 May 2005). 
Smith, S.R., Jr., and Whalley, R.D.B. 2002. A model for expanded 
use of native grasses. Native Plants J. 3: 38^-9. 
Smoliak, S., and Johnson, A. 1980. Elbee northern wheatgrass. 
Can. J. Plant Sci. 60: 1473-1475. 
Smoliak, S., and Johnson, A. 1983. Walsh western wheatgrass. 
Can. J. Plant Sci. 63: 759-761. 
Sokal, R.R., and Michener, CD. 1958. A statistical method for 
evaluating systematic relationships. Univ. Kans. Sci. Bull. 38: 
1409-1438. 
Vogel, K.P., and Pedersen, J.F. 1993. Breeding systems for cross- 
pollinated perennial grasses. Plant Breed. Rev. 11: 251-274. 
Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., 
Homes, M., et al. 1995. AFLP: a new technique for DNA fin- 
gerprinting. Nucleic Acids Res. 23: 4407^1414. 
Wagnon, H.K. 1952. A revision of the genus Bromus, section Bro- 
mopsis, of North America. Brittonia, 7: 415^-80. 
2005 NRC Canada