Broad-Scale Integrity and Local Divergence in the Fiddlehead Fern Matteuccia struthiopteris (L.) Todaro (Onocleaceae)

2011·9.7 MB·English

by  Daniel M Koenemann

and
Broad-Scale
Integrity Local Divergence
in the
Fiddlehead Fern
Matteuccia
struthiopteris
(L.)
Todaro
(Onocleaceae)
M. Koenemann
Daniel
Rancho Santa Ana Botanic Garden, 1500 North College Avenue, Claremont, CA 91711-3157,
Jacqueline A. Maisonpierre
University Vermont, Department
of of Plant Biology, Jeffords Hall, 63 Carrigan Drive. Burlington.
VT
05405, e-mail: [email protected]
David Barrington
S.
VT
05405, [email protected]
e-mail:
Matteuccia
Todaro
struthiopteris (Onocleaceae) has present-day
(L.) a
much
distribution across of the north-temperate and
boreal regions of the world.
commonly known
It is as the fiddlehead fern in the northeast of North America
because of the resemblance namesake
of its elegant croziers to their (Kato, 1993).
Matteuccia struthiopteris found most commonly on banks and
is river flood
However,
plains. the species can also be found in uplands an5rwhere there are rich
Though we
alluvial soils (Smith, 1993). robust in full sun, have found that the
common
more
ferns are in either full or partial shade, sheltered by broadleaf
deciduous trees. They often form large colonies on wooded floodplains via
spread by The
prolific stolons. fronds are dimorphic; large (0.3 to greater than
m)
produced
1.5 sterile fronds in the spring are followed by smaller (0.3 to 0.7 m)
fertile fronds produced in late summer, which have their sporangia rolled up
inside of the pinnules (Prange and von Aderkas, The
1985). sterile fronds are
oblanceolate, tapering gradually at the base and abruptly at the tip; the widest part
of the frond subapical (Prange and von Aderkas,
is 1985). Sterile fronds present
two morphological we downy and downy
variants that call smooth. In the variant
young
the petioles of fronds are covered with a minute than mm),
(less 0.5
AMERICAN FERN VOLUME NUMBER
JOURNAL:
101 4 (2011)
non-laminar indument. smooth
colorless, In the variant the petiole lacks this
downy
indument. This indument summer,
dissipates in early so presence
its
cannot be scored later in the growing season.
much
Matteuccia struthiopteris distributed throughout of the northern
is
hemisphere
in boreal deciduous forests (Lloyd, 1971); collected as a seasonal
it is
New
edible green in England and the eastern provinces of Canada (von Aderkas,
1984), as well as in Japan (Miyazawa et ah, 2007). Fiddlehead fern collection
is
documented
American
for several Native groups, including primarily the
New New
Abenaki Indians of England and Malecite of Brunswick. European
colonists learned of this harvest practice from these native groups on their arrival
North America and
in collected fiddleheads as part of a subsistence diet (von
we
Aderkas, 1984). In Vermont, have seen a recent explosion of interest in buying
and
preparing fiddleheads with the rapid expansion of the local-foods movement.
In the last three years, landowners have begun to post their lands with no-
harvesting signs for the time.
first
known
Little is about the impact of fiddlehead harvesters on the demography or
genetic structure of Matteuccia struthiopteris, as their activity not regulated.
is
known
However,
it is that the species' health can be significantly affected by the
removal of young fronds. Bergerson and Lapointe conducted
(2001) a five-year
study in Quebec, Canada, investigating the impacts of harvesting on Matteuccia
struthiopteris. They found that harvesters should collect no more than half the
croziers produced by a shoot, as removal of all shoots in a single year impaired
carbohydrate accumulation and production subsequent
crozier in years.
With
the advent of molecular techniques, there has been substantial
improvement
in our understanding of the geographic distribution of genetic
variation in plant populations. In a historical context, these distributional
now
commonly
patterns are used to interpret the recent distributional history of
(Donoghue make
species et al, 2001; Hewitt, 2000). Analyses of these patterns the
we
basic assumption, an assumption that will maintain, that populations near
Pleistocene having remained
refugia, established for the longest period of time,
more
will display greater genetic diversity than peripheral populations. This
was
concept proposed model
first in the stepping-stone of population structure
developed by Kimura and Weiss Building on models
(1964). this idea, dispersal
constructed by Ibrahim and Hewitt
et al. (1996) (1996) predicted that as a species
expands from refugium newly
the established populations contain
its less
diversity than the parent population. The ideas of Kimura and Weiss
(1964),
Ibrahim and Hewitt were confirmed subsequent by
et (1996), (1996) in studies
al.
Dufresne and Hebert (1997) and most of those reviewed by Harrington and Paris
(2007).
On
the other hand, genetic variafion in populations relates to local evolutionary
and and
history (both selection population
drift) factors relating to size (large
and
central small peripheral populations can evidence
different patterns of
genetic diversity). In northeastern North America for instance, the distribution of
Cypripedium
genetic diversity in parviflorum Salisb. appears to be the result of
and
genetic drift related to differences in population size not historical factors
and
(Wallace Case, 2000). Given the great differences in population size of
KOENEMANN
ET
M. STRUTHIOPTERIS HOLOCENE DYNAMICS
AL.:
215
Matteuccia with
elevation (very large in alluvial forests low in watersheds and
small and wedands
isolated in high and
at elevations], history recent evolution
must
both be
considered.
Our
interest in the historical biogeography Matteuccia
of
struthiopteris led us
to
assess genetic diversity of the fiddlehead two we AFLP
fern used
at scales.
First,
document
analysis
to the distribution of genetic diversity within an array of
fiddlehead fern to assess variation within and among We
populations Vermont.
in
expected one two
of patterns emerge:
to concentration
1) of variation in either
lowland
large or small highland populations implying
recent selecdon
drift or or
2) a regional pattern of decreasing diversity implying We
a
historical cause. also
explored the pattern of variation in the downy versus smooth
variants in these
populations with we
similar expectations about Second,
pattern.
set this local
variation in the context of a less detailed view of the global genetic structure
for
DNA
the fiddlehead fern based on and
chloroplast nuclear
sequences.
In this
we
study sought human
to assess potential impact on
genetic
diversity, as
now
harvesting
of these
ferns is intensive in the large populations in lower-
elevation
river valleys of the
region.
Materials and Methods
Taxon A
sampling.—
total of 83 accessions representing four onocleaceous
taxa
was
included
in this study. All but three of these accessions were Matteuccia
struthiopteris. Unless noted
otherwise, each samples were
at site collected as
Up
follows. to ten complete fronds were each from
collected,
a different shoot, hi
an
avoid
effort to collecting multiple ramets was
of single genets, each frond taken
m
minimum
at a distance of 2 from any other sample. The maximum was
distance
m
from any Downy
5 other sample. individuals were when
deliberately included
encountered
in a population, and each population was downy
scored
as
abundant,
rare, or absent. Material collected from each
individual included
a
DNA
sample
for analysis stored -80 C and
in silica gel at a dried voucher
specimen, both from same
the
frond.
A
total of 60 individuals of Matteuccia was
struthiopteris collected across three
watersheds Vermont:
in the Passumpsic, Mettawee, and Winooski
rivers
(Fig.
1).
The
Passumpsic
River
a fributary of the Connecticut
is River, located
entirely
within Kingdom
the Northeast
of Vermont. Three were made
collections along
the
Passumpsic,
the highest Newark
in elevation in at 358 m, the middle in East Burke
m
m, and Bamet
at 253 the lowest in 156 (See Appendix
at for accession data
1
and voucher
information).
The Mettawee
River
the shortest of the three systems from which
is river
plants
were collected. One collection was made along Mettawee
the River
in Granville,
m
NY
173 (Appendix
at
1).
The
Winooski
River
is the largest of the three watersheds surveyed. The
river
an
drains area of the northern section of the Green Mountains;
a large tributary
is
it
Lake Champlain. Three were made
to collections along the The highest
river. in
elevation is in Cabot at 442 m, the next in Plainfield 238 m, and
at the lowest in
m
Richmond
87 (Appendix
at
1).
AMERICAN FERN VOLUME NUMBER
JOURNAL:
101 4
(2011]
The
remaining 23 accessions comprised two species of the Matteuccia segregate
genus Pentarhizidium intermedium Hayata and
(C.Chr.) P. orientale (Hook.)
[P.
Hayata), the out-group taxon Onoclea sensibilis and 20 additional Matteuccia
L.,
New
struthiopteris accessions including 15 from the World (Vermont, Maine,
New
Quebec, Brunswick, Michigan, British Columbia, and Alaska) and four from
the Old World (China, Japan, and Sweden). Accessions from China and Sweden
were provided by correspondents without vouchers or information on collection
method (Appendix
1).
—
DNA DNA
extraction. Total was extracted from approximately 0.05 of frond
g
CTAB
tissue following Dempster et (1999), which based on the protocol of
al. is
KOENEMANN
ET STRUTHIOPTERIS HOLOCENE DYNAMICS
M.
AL.:
Doyle and Doyle (1987). The following slight modifications were made in order
acquire samples high and
to of the quality concentration required the
for
AFLP
technique. Homogenized frond material was allowed incubate
to
at
CTAB
C
55 24 hours
for in buffer. After the addition of isopropanol the
DNA
were -80
nucleic acids allowed to precipitate for one hour at C. was
70%
pelleted using centrifugation, then cleaned once ethanol wash and
in a
95%
once in a ethanol wash. The pellets were dried via tube inversion or
New DNA
Speedvac (Genevac Inc, Gardiner, York, USA). After the was
resuspended in 50 of ddHsO the concentration was quantified using
|iL a
NanoDroplOOO
(Thermo Waltham, Massachusetts, USA). Samples
Scientific,
with
concentrations lower than 100 were and
ng/|iL discarded re-extracted
following
Sigel (2008).
PCR
amplification— Scieening of seven chloroplast markers and one nuclear
marker two
for variability yielded variable markers sequencing, the highly
for
labile cp marker psbA-trnH spacer [psbA-trnH, see Kress 2005) and the
et
al.,
PgiC region including most of intron 15 and a portion of exon 16 [PgiC 15-16),
work
labile in with Dryopteridaceae in our lab (Sigel, 2008). psbA-trnH pnmeTs
Sang
are those of et al. (1997) with modification for use on ferns (psbA-trnHFor:
GTTATGCATGAACGTAATGCTC
CGCGCATGGTG-
5' psbA-trnHRev: 5'
3',
GATTCACAATCC
The
primers PgiC 15-16 were modified study
3'). for for this
from those developed in our lab for use on Polystichum (Dryopteridaceae),
TTTGCTCCTCACATT-
following Ishikawa They PgiGlSFM
et al. (2002). are (5'
CAACA
GTTGTCCATTAGTTCCAGGTTCCCC
and PgiClORM
These
3'), (5'
3').
regions were amplified by the polymerase chain reaction (PCR) method with
all
thermal cycling done using either a model TC-312 or T-3000 thermal cycler
New
(Techne, Burlington, USA). Reactions were completed 24
Jersey, in
^iL
aliquots with the following reaction components: for psbA-TrnH: 100-200 ng of
DNA,
0.1 ^imol/L of each primer, X ExTaq Buffer (TaKaRa, Madison, Wisconsin,
1
USA), 200 ^mol/L of each dNTP, 200 ^mol/L MgCl2, 200 BSA, and
of [imol/L of
U
ExTaq DNA,
0.625 polymerase (TaKaRa); PgiC 15-16: 100-200 ng
for of
IX
0.1 ^mol/L of each primer, ExTaq Buffer, 400 ^mol/L of each dNTP, 400 ^mol/
U
L BSA, and Ex Taq The
of 0.625 polymerase. reaction conditions were
as
follows for psbA-trnH: initial denaturation was for 1 min at 95 C; followed by 30
cycles of 1 min at 95 C, 1 min at 50 C, and 4 min at 65 C; with a final extension of
C
72 for 5 minutes. For PgiC 15-16: initial denaturation was for 3 min at 95 C;
followed by three cycles of min 94 min 56 and min 72
1 at C, 1 at C, 2 at C;
followed by min min and min
three cycles of 1 at 94 C, 1 at 53 C, 2 at 72 C;
followed by 34 cycles of 45 sec at 94 C, 45 sec at 50 C, and 90 sec at 72 C; with a
C PCR
final extension of 72 for 8 min. Before sequencing the products were
ExoSAP-IT (USB
purified using Corp., Cleveland, Ohio, USA).
Agarose and
gel electrophoresis extraction.—Priov to sequencing,
all
2%
amplified samples were visualized on agarose with ethidium bromide
gels
mL
EDTA
(1.0 agarose, 50 Tris Borate buffer, 2.5 ethidium bromide). two or
g [iL If
more bands were separated during band
electrophoresis, the of the appropriate
DNA
marker was and was
length excised the amplified extracted using the
QIAquick
Gel Extraction Kit (Qiagen Inc., Valencia, California, USA). Before
AMERICAN FERN VOLUME NUMBER
JOURNAL:
101
4 (2011)
sequencing the extraction products were purified using ExoSAP-IT (USB
Corp.,
Cleveland, Ohio, USA).
Sequencing.—
Direct sequencing of both markers was undertaken using the ABI
BigDye Terminator Cycle Sequence Ready
Reaction
Kit 31 (Perkin-Elmer/
v.
Applied Biosystems, Foster USA). The employs
City, California, the following
kit
thermal-cycling protocol: initial denaturation was for 5 min 80 "C followed by
at
C C
30 cycles of 10 sec at 96 and five sec at 50 with a final extension at 50^C for
An
ABI
4 min. Prism 3130x1 automated sequencer was used
resolve the
to
DNA
sequencing products (Vermont Cancer Center Analysis
Burlington,
Facility,
Vermont
USA).
—
Raw
Sequence
alignment. forward and reverse sequences were assembled
for
each sample and ambiguous were
bases from
corrected inspection of the
chromatograms; Sequencher and
3.1.1 (Nishimura, 2000) BioEdit
v. 7.0.9
v.
(Hall, 1999) were used sequence and consensus-sequence
for editing construc-
Consensus sequences were
tion. first aligned using ClustalX (Larkin et al, 2007),
then manually improved using MacClade (Maddison and Maddison,
v.4.08
2005).
Phylogenetic ana/ysis.—Separate analyses were conducted each
data
for set
was
that generated [psbA-trnH, PgiC 15-16) and the two combined using
Maximum
we PAUP*
Parsimony. For used
the analysis version 4.0bl0 (Swofford,
A
2001) with all characters treated as unordered with equal weights. heuristic
was
search performed using 10,000 replicates and random taxon addition with
ten trees held at the tree-bisection-reconnection (TBR) branch swapping step with
A
maximum
each sequence was
addition. of ten saved each
trees at step,
ACCTRAN
MulTrees
option with
on, character-state optimization. Bootstrapping
was
performed 1000 TBR
for replicates using simple taxon addition, branch
swapping, and MulTrees we
the option For
on. this analysis, retained a single
We
representative accession
for sets of identical accessions
(see Fig. 2 for details).
considered bootstrap percentages greater than 68 and 95 be moderate and
to
strong levels of clade support respectively following and Barrington
Driscoll
We
were
(2007). All trees rooted using Onoclea sensibilis as the outgroup. report
only combined
the analysis: separate analyses were congruent with but
less
resolved than the combined
analysis.
AFLP
AFLP
protocol.—The was
protocol adapted from Vos with
et (1995)
al.
additional modifications as suggested by the Wolf Lab Utah University
at (Wolf,
Gastony Lab
2000), the at Indiana University (Nakazato et 2006), previous
al.,
studies in the Barrington Lab (Sigel, 2008), as well as modifications specific to this
study. Several procedures were performed minimize genotyping
to error in the
A
final data. previously typed sample was included each round
in of
amplification comparison between Ten
for different rounds. percent of the
samples were between two and
replicated four times to assess the error within
samples. Those samples with manifest such
errors in amplification, as greatly
reduced numbers, were and
allele discarded re-amplified following
Sigel
(2008).
DNA
Restriction of the genomic was performed using
the restriction
endonucleases EcoRl and Msel (New England
Biolabs, Ipswich, Massachusetts,
USA).
Ligation of the sticky end-adaptors (Invitrogen, Carlsbad, California, USA)
was
to the restriction sites also achieved in this reaction following the Wolf
initial
KOENEMANN
ET STRUTHIOPTERIS HOLOCENE DYNAMICS
M.
AL.:
(2000) protocol. Modifications from the Wolf Lab protocol (Wolf, 2000) are as
follows: 20 of the enzyme master mix were used avoid volumes
pipetting
to
under 0.4 |iL. The master mix had the following components: 0.02 of Msel,
|aL
0.05 i^L of EcoRl, 0.025 [iL of T4 Ligase, and 0.655 |iL ddHsO were used per 1 ^iL of
master-mix Each volume
total product. reaction totaled 11 each containing
fiL,
DNA
100 ng of suspended in 5.5 ddHaO. Samples were incubated two
^iL for
hours 37'C thermal
at in a cycler. In preparation for pre-selective amplification
3 |iL of the restriction/ligation product were diluted with 24.5 TE
0.1 buffer.
round PGR, was
In the of preselective amplification achieved with primers
first
that complement the EcoRl and Msel adaptors plus one additional nucleotide
i.e.,
Msel+C GAT GAG TCG TGA GTA AC GAG TGC GTA CCA
and EcoKl+A
(5'
3') (5'
ATT CA
Each
3'). preselective reaction totaled 25 comprising the following:
|iL,
ExTaq
3 |iL diluted restriction/ligation product, 2.5 |iL Buffer (TaKaRa), 0.1 Ex
ixL
TaqDNA
50mM
polymerase (TaKaRa), 2.5mM dNTPs,
0.75 of MgCla, of
1.2
ddHsO, and
16.45 |iL 0.5 |iL of a 10 |iM solution of each of the primers. Reaction
samples were min
initially denatured for 2 at 72 C, followed by 30 cycles 94"G
of:
C
for 30 sec, 56 for 30 sec and 72 C for 2 min, ending with 30 min at 6 C. Fifteen
and
a half of the preselective amplification product were diluted with 100 of
TE
The
0.1 buffer. quality of the undiluted preselective and restriction/ligation
was
product visualized on a 1.5% TBE-agarose
gel.
In the second round of PGR, selective amplification was achieved using the
Msel+CAG NED
and fooRI+AAC
(fluorophore [yellow] primers); each selective
primer has an sequence
identical to corresponding preselective primer but
its
with an two The
additional EcoRl primer
bases. fluorescently tagged
is for
on ABI
visualization the 3100-Avant Genetic Analyzer (Applied Biosystems,
Foster City, California, USA). Selective reactions were set up in 18 aliquots
(jL
DNA
comprising
the following: 3.6 of the diluted preselective product,
|iL 1.8
^iL
DNA
Ex Taq Ex Taq
Buffer, 1.8 polymerase, MgCla, 0.864 dNTPs,
|iL 0.5
^iL
0.4mM
1.44 of the Msel primer 0.08mM
|iL selective (Invitrogen), 3.66 of the
EcoRI
selective primer (Invitrogen), 0.144 BSA, and 5.872 ddH20. Reactions
|iL
were initially denatured at 94"C for 2 min, followed by 13 cycles of 94 C for 30 sec,
65 C for 30 sec (reduced by 0.7 C per cycle), 72 C for 2 min, and 24 cycles of 94 C
G C
for 30 min, 56 for 30 sec, 72 for 2 min, with a final hold at 72 C for 30 min.
Selective amplification products were run on the 4-capillary ABI 3100-Avant
DNA
Genetic Analyzer the Vermont Cancer Center Analysis
at Facility at the
University of Vermont.
ALFP
data scoring.—The most AFLP
challenging aspects of analysis are data
scoring and analysis (Bonin Bonin Meudt and
et al, 2004; et al, 2007; Clarke,
Pompanon
Vekemans
2007; et 2005; 2002; Whitlock
al., et et 2008). In
al., al.,
AFLP
principle, the challenge to data scoring producing a set of binary
is
phenotypes
(termed loci) for each individual from the presence/absence of the
DNA
fragments retrieved, while the same time excluding experimental
at
We
artifacts. used the loci exclusion thresholds and phenotype exclusion
thresholds outlined in Whitlock et al. (2008) to reduce the likelihood of scoring
while maintaining maximal number Appendix
artifacts a of informative markers.
AMERICAN FERN VOLUME NUMBER
220 JOURNAL:
101 4
(2011)
B AFLP
of Sigel (2008) provides a detailed account of analysis problems and
solutions our
in
lab.
AFLP
Initial fingerprints were determined with GeneMapper (Applied
3.7
v.
Biosystems). Peak-height data for each individual were exported in tab-delineated
form to an Excel worksheet. The top ten percent of the peak heights was trimmed
remove The mean
to false flares in the recording instruments. of the trimmed data
was
calculated for each locus across all individuals as well as for each individual
across all its loci. These means were used as a basis for defining an array of
candidate loci-exclusion and phenotype-calling Using trimmed
thresholds. the
mean
rather than an arbitrary value of 100 relative fluorescence units as
(rfu)
suggested by GeneMapper (Applied Biosystems) more accurate
as better
is
it
We
represents the specific data set under study (Sigel, 2008). generated 20 binary
with an and
datasets array of loci-exclusion phenotype-calling threshold
combinations. The mismatch (which number
error rate represents the of
individuals with band was
inconsistent data in replicate amplifications)
calculated for each generated dataset. The optimum threshold was determined
by maximum
number
selecting the generated dataset that maintained the of
informative while mismatch below
loci retaining a error rate five percent (Bonin
Binary phenotype optimum
et al, 2004). datasets generated using the identified
thresholds were used
in the data analysis.
—
Data and The maximum
analysis interpretation. binary dataset with the
number
of informative with an below was used
loci error rate five percent as a
GENALEX
basis for data analyses. version 6.2 (Peakall and Smouse, 2006) was
A
used
for all statistical analyses of the data. pairwise, individual-by-individual
was
genetic-distance matrix generated. Each value in the matrix a of
tally
is
differences between two genetic profiles (Peakall and Smouse, 2006, Appendix
1).
As
an initial probe of the data the genetic-distance matrix was used do a
to
We
PGA
principal-components analysis (PGA). explored the by visualizing
individual plants on plots of pairs of principal components accounting
for
substantial variance in the data to search for clusters of individuals according
to
and
population, watershed,
elevation.
To
assess patterns of genetic diversity across the seven Vermont populations
AFLP
and were
elevations the data subjected to three calculations: the average
genetic distance between individuals, the average heterozygosity, and the
total
number
of loci present.
Results
—
Vermont downy was
Field observations. In the absent in the high-
trait
elevation population in the Winooski watershed and in both the high and middle
populations in the Passumpsic watersheds. The trait was rare in the middle-
elevation population in the Winooski watershed. Both watersheds had an
downy
abundance The downy
of individuals in the low-elevation population.
was
character trait not scorable for individuals in the Mettawee population, as
it
was New
collected too late in the season. In the Maritimes, three low-elevation
New
Brunswick
populations one near Gampbellton Brunswick, one
(the the the
at
KOENEMANN
ET STRUTHIOPTERIS HOLOCENE DYNAMICS
M.
AL.:
intersection of Route 2 Canada and the Canaan River, and the one just west of
downy
Fredericton) included
plants.
—
Sequence The and
characteristics. aligned concatenated sequence psbA-
for
trnH and PgiC 15-16
yielded a matrix of 798 characters, including
five indels.
psM-frnf/ comprised 439 of these characters including two of the indels. Within
Matteuccia, 59 were which
characters (7.4%) variable, of 36 including four indels
were parsimony Twenty
(4.5%) informative. characters separated Onoclea
(the
outgroup) from the study group. In PgiC, direct sequencing yielded largely
unambiguous
sequences, presumably because most were homozygous
plants
for
An
all nucleotides. informative exception was our sequence Onoclea, which
for
contained an extensive region of double peaks at the 3' end. The pattern of
double peaks was
in this region consistent with the interpretation that they were
the result of the insertion of one nucleotide in one of the two homologs;
for
we
was
analysis retained the allele that identical to the rest of the study
set.
Phylogenetic analysis—The
phylogenetic analysis conducted using Maxi-
mum
Parsimony
(MP) yielded one The
shortest tree consisting of 53 steps.
=
analysis (Fig. 2) retrieved a strongly supported clade (BS 100) comprising
Asian endemics
the Pentarhizidium intermedium and
orientale be
P. to sister
=
monophyletic
to a (BS 100) M. struthiopteris. Within our study species,
Eurasian and American M.
struthiopteris clades were moderately well
supported (BS value each Within each two M.
for is 86). of the struthiopteris
clades all accessions were unresolved, with the exception of two of the
= An
accessions from eastern Canada, which were each other (BS
sister to 100).
additional analysis, including the Alaska accession was received while
that
was
this contribution in review, yielded similar but with the Old
results,
World 66% new
clade bootstrap support reduced to and a clade comprising the
Old World and 65%
Alaska with
retrieved bootstrap support.
New
Looking
at individual characters for Matteuccia struthiopteris. World
and Eurasian accessions by two them
differed three substitutions (0.3%), of
New
synapomorphies Old World and one World
for the plants for the plants.
With
the addition of the Alaska accession, there was one synapomorphy
unique to the Old World plants, one shared by the Old World and Alaska
New
accessions, and one by the World plants excluding the Alaska accession.
The
nearly complete lack of resolution within M. due
struthiopteris to
is
absence synapomorphies
of rather than character incongruence.
AFLP AFLP
analysis—
analysis of the sixty Matteuccia struthiopteris samples
Msel+CAG fcoRI+AAG
using the primer 254
pair resulted in distinguishable
loci.
A
75%
phenotype-calling threshold of (92.51 and loci-exclusion threshold of
rfu)
85%
(104.85 rfu) were used to construct the binary data matrix from the raw peak-
height data. This combination of thresholds applied to the raw data resulted in a
4.98% mismatch
between
error rate replicate samples.
The percent polymorphic loci across populations was 86.22 while the
all
polymorphic
percent loci within the individual populations ranged from 33.36
in the Barnet population (middle elevation population in the Passumpsic
watershed) to 61.81 in the Plainfield population (middle elevation population
Winooski
in the watershed, Table
1).
AMERICAN FERN VOLUME NUMB
JOURNAL:
101
Maine
(Dac3)
Michigan
(RezOI)
Columbia
(Sea1071)
British
New
Brunswick (DMK4109)
Vermont
(JAM001)
Quebec (DMK1109)
New
Brunswick (DMK3109)
Quebec (DMK0109)
Vermont (MAM0709)
China
(CB15)
Sweden
(AHR001)
Hokkaido (MOC001)
Hokkaido (MOC002)
Pentarhizidium intermedium
(2D7)
Pentarhizidium orientale (DL47)
(DMK10109)
Bootstrap analysis of A based on psbA-tmH and PgiC 15-16
DMK
DMK
4109 and JAMOOl), (same
s as 0109).
The
principal-components
analysis revealed substantial geographic clustering
The
of genetically related individuals. three components retrieved from the
first
accounted 75.52%
analysis for of the variance; they represented 32.72%, 29.95%,
and 12.85%
of the total variance in the data respectively. Principal components
1
and 3 were most powerful in portraying the relationships among individuals and
populations Most of the Passumpsic-watershed
(Fig. plants lay in a tight
3).
cluster; they were largely separated from the Winooski-watershed and Mettawee-
watershed on component
plants principal In confrast the Winooski-watershed
1.