Plant communities in broad-leaved deciduous forests in Hordaland county, Western Norway


Aarrestad, P.A.

Nordic Journal of Botany 20(4): 449-466

2000


The vegetation of broad-leaved deciduous forests in Hordaland, Western Norway is described with regards to species composition and structure. The investigation is based on phytosociological analyses of native forest stands dominated by Corylus avellana, Fraxinus excelsior, Ulmus glabra, or Tilia cordata. Two-way-indicator analysis (TWINSPAN) and correspondence analysis (CA) are used to distinguish different vegetation types and assess possible gradients in the vegetation data. The vegetation types are discussed in relation to equivalent forest communities and to syntaxonomical units. Two forest types are described at the first hierarchical TWINSPAN level: 1) A hygrophilous and slightly thermophilous Fraxinus excelsior-Cirriphyllum piliferum forest characterised by an open tree canopy, a dense field layer of tall herbs and ferns, and a high cover of bryophytes. 2) A thermophilous and less hygrophilous Corylus avellanα-Brachypodium sylvaticum forest characterised by a dense tree layer, a more open field layer with larger elements of small herbs, and a somewhat lower bryophyte cover. The CA analysis clearly separates the samples from the first TWINSPAN division along the first ordination axis. Five forest types have been described at the second hierarchical level, mainly associated with differences in mesotrophic and eutrophic species, but there is clearly a gradient structure in the species composition between these plant communities. In relation to syntaxonomy, the first TWINSPAN division supports the separation of the west Norwegian broad-leaved deciduous forests into a hygrophilous plant community, the Eurhynchio-Fraxinetum (Blom 1982) Ovstedal 1985 and a drier and more thermophilous community, the Primulo-Ulmetum (Blom 1982) Ovstedal 1985.

Nordic
Journal
of
Botany
Plant
communities
in
broad-leaved
deciduous
forests
in
Hordaland
county,
Western
Norway
Per
Arild
Aarrestad
Aarrestad,
P.
A.
2000.
Plant
communities
in
broad-leaved
deciduous
forests
in
Hordaland
county,
Western
Norway.
Nord.
J.
Bot.
20:
449-466.
Copenhagen.
ISSN
0107-055X.
The
vegetation
of
broad-leaved
deciduous
forests
in
Hordaland,
Western
Norway
is
described
with
regards
to
species
composition
and
structure.
The
investigation
is
based
on
phytosociological
analyses
of
native
forest
stands
dominated
by
Corylus
avellana,
Fraxinus
excelsior,
Ulmus
glabra,
or
Tilia
cordata.
Two
-way
-indicator
analysis
(TWINSPAN)
and
correspondence
analysis
(CA)
are
used
to
distinguish
different
vegetation
types
and
assess
possible
gradients
in
the
vegetation
data.
The
vegetation
types
are
discussed
in
relation
to
equivalent
forest
communities
and
to
syntaxonomical
units.
Two
forest
types
are
described
at
the
first
hierarchical
TWINSPAN
level:
1)
A
hygrophilous
and
slightly
thermophilous
Fraxinus
excelsior-
Cirriphyllum
piliferum
forest
characterised
by
an
open
tree
canopy,
a
dense
field
layer
of
tall
herbs
and
ferns,
and
a
high
cover
of
bryophytes.
2)
A
thermophilous
and
less
hygrophilous
Corylus
avellana-Brachypodium
sylvaticum
forest
characterised
by
a
dense
tree
layer,
a
more
open
field
layer
with
larger
elements
of
small
herbs,
and
a
somewhat
lower
bryophyte
cover.
The
CA
analysis
clearly
separates
the
samples
from
the
first
TWINSPAN
division
along
the
first
ordination
axis.
Five
forest
types
have
been
described
at
the
second
hierarchical
level,
mainly
associated
with
differences
in
mesotrophic
and
eutrophic
species,
but
there
is
clearly
a
gradient
structure
in
the
species
composition
between
these
plant
communities.
In
relation
to
syntaxonomy,
the
first
TWINSPAN
division
supports
the
separation
of
the
west
Norwegian
broad-leaved
deciduous
forests
into
a
hygrophilous
plant
community,
the
Eurhynchio-Fraxinetum
(Blom
1982)
Ovstedal
1985
and
a
drier and
more
thermophilous
community,
the
Primulo-Ulmetum
(Blom
1982)
Ovstedal
1985.
P
A.
Aarrestad,
Norwegian
Institute
for
Nature
Research,
Tungasletta
2,
N-7485
Trondheim,
Norway.
E-mail:
per.a.aarrestad@ninatrd.ninaniku.no.
Introduction
This
study
deals
with
the
so-called
'noble'
deciduous
forests,
in
Central
Europe
called
"Edellaubwalder"
(Ellenberg
1986)
and
in
Norwegian
called
"edel-
lauvskog",
where
the
term
refers
to
forests
dominated
by
Acer
platanoides,
Alnus
glutinosa,
Fraxinus
ex-
celsior,
Fagus
sylvatica,
Quercus
spp.,
Tilia
cordata,
and
Ulmus
glabra
(Korsmo
1974).
In
Western
Norway
Accepted
15-6-2000
Nord..1.
Bot.
20(4)
2000
the
broad-leaved
deciduous
forests
are
located
at
their
northern
distribution
limits
in
Europe.
In
this
area
they
have
formerly
been
more
widespread,
especially
in
the
warm
period
of
the
Atlantic-
and
early
Subboreal
time
(cf.
Kvamme
1988;
Presch-Danielsen
1996).
However,
changes
in
climate
and
conversion
of
large
areas
into
agricultural
land
have
reduced
the
areas
of
therm-
ophilous
forests.
The
remaining
stands
now
occur
on
non
-arable
land,
mainly
in
areas
of
steep
hill
slopes
449
with
more
or
less
nutrient
-rich
soils
and
a
favourable
local
climate
(Austad
&
Skogen
1990).
Several
of
these
stands
have
previously
been
influenced
by
different
kinds
of
management
such
as
coppicing,
thinning,
grazing
and
fodder
production
involving
both
the
tree
and
field
layers
(e.g.
Ve
1940,
1941;
Lea
1984;
Austad
et
al.
1985;
Austad
&
Skogen
1990;
Austad
&
Losvik
1998).
The
investigated
forests
in
this
study
show
some
signs
of
earlier
management,
but
are
today
only
weakly
utilised
or
are
not
utilised
at
all,
and
can
thus
be
considered
as
more
br
less
native
or
semi
-natural
forests.
They
are
characterised
by
several
nutrient
-
demanding,
frost
-sensitive,
and
thermophilous
species
and
the
dominant
trees
are
Fraxinus
excelsior,
Corylus
avellana,
Tilia
cordata,
and
Ulmus
glabra.
Acer
pseudoplatanus
is
also
frequently
occuring.
However,
this
tree
was
originally
an
antropogenic
species
on
the
west
coast
of
Norway
(Fremstad
&
Elven
1996),
but
is
today naturalised
in
several
of
the
broad-leaved
deciduous
forests.
Even
though
woodland
species
are
the
most
important;
there
is
a
distinct
element
of
heliophilous
species,
both
margin
species
and
cultural
indicators.
This
is
probably
due
to
former
management
and
to
the
abrupt
topography
causing
natural
openings
inside
the
woodlands.
Dense
stands
of
Alnus
glutinosa
are
found
on
waterlogged
soils
in
swamps,
along
streams
and
brooks
and
on
fl
ushed
hillsides
(Fremstad
1983),
but
they
are
not
considered
in
this
study,
neither
are
the
sparse
occurrences
of
oligotrophic
oak
forests.
The
vegetation
of
broad-leaved
deciduous
forests
in
Hordaland
has
been
investigated
for
several
purposes.
Korsmo
(1975)
surveyed
localities
in
connection
with
the
national
plan
for
protection
of
forests
in
Norway.
Several
studies,
mainly
of
a
phytosociological
character,
have
been
carried
out
in
limited
areas,
both
in
native
and
cultivated
forests
(Kielland-Lund
1974;
Balle
1978;
Losvik
1978;
Blom
1980;
Fottland
1980;
Austad
&
Skogen
1988).
Aarrestad
(1985)
described
soils
in
meso-
and
eutrophic
forests
that
were
related
to
releves
described
in
the
thesis
work
of
Blom
(1982)
and
Fottland
(1982).
Rosberg
&
Ovstedal
(1987)
studied
the
phytosociology
and
soils
of
Corylus
avellana
coppices
in
the
westernmost
parts
of
Hordaland,
and
Odland
(1992)
has
investigated
stands
of
Matteuccia
struthiopteris.
More
recently,
a
study
has
been
carried
out
on
the
representativeness,
biological
diversity
and
species
rarity
in
a
gradient
from
west
to
east
in
Hordaland
(Swtersdal
&
Birks
1993;
Seetersdal
et
al.1993;
Szetersdal
1994).
This
study
is
a
part
of
an
ecological
investigation
of
the
mesotrophic
and
eutrophic
broad-leaved
deciduous
hill
forests
in
Hordaland,
where
vegetation
analyses
(phytosociological
releves)
and
environmental
variables
(including
soil
samples)
have
been
collected
from
the
same
plots
in
order
to
investigate
vegetation
and
environment
relationships.
The
aim
of
the
present
paper
is
to
describe
the
fl
oristic
composition
of
the
vegetation,
as
a
basis
for
further
studies
on
the
ecology
of
these
forests.
This
will
be
done
by
numerical
classification
of
the
releves
in
order
to
identify
vegetation
types,
and
indirect
gradient
analysis
will
be
used
to
detect
gradients
in
the
species
composition
and
to
evaluate
the
separation
of
the
vegetation
types.
Finally
the
described
plant
communities
will
be
compared
with
similar
forest
communities
in
Scandinavia
and
in
Great
Britain
and
related
to
described
phytosociological
taxa.
Investigated
area
The
investigation
covers
46
localities
in
the
local
councils
of
Bergen,
Osteroy,
Os,
Fusa,
Samnanger,
Stord,
Tysnes,
Kvinnherad,
Kvam,
Granvin,
and
Ulvik
in
Hordaland
county
(Fig.
1).
A
detailed
list
of
the
investigated
localities
is
available
for
readers
on
request.
The
investigated
area
belongs
to
the
boreo-nemoral
region,
a
transitional
zone
between
the
nemoral
deciduous
forests
to
the
south
and
the
boreal
region
to
the
north,
dominated
by
conifers,
birch,
and
grey
alder
forests
(Dahl
et
al.
1986;
Moen
1987,
1998).
The
bedrock
consists
of
sedimentary,
volcanic
and
plutonic
rock
types
of
Cambro-Silurian
age
and
rocks
of
Precambrian
age,
all
more
or
less
metamorphosed
during
the
Caledonian
deformation
(Askvik
1976).
Most
of
the
rich
deciduous
hill
forests
are
located
on
south-east-
to west
-facing
steep
slopes
below
vertical
rocky
walls
with
rock
-fall
material
(talus)
or
scree
material
as
the
underlying
substratum.
The
regional
climate
is
oceanic
with
mild
winters,
relatively
cool
summers
and
high
annual
precipitation.
Within
the
investigated
area
there is
a
distinct
climatic
gradient
from
the
coast
to
the
relatively
continental
areas
in
inner
Hardanger,
both
in
precipitation
(Forland
1979,
1993)
and
temperature
(Aune
1993).
Material
and
methods
The
fieldwork
was
carried
out
during
the
summer
and
early
autumn
from
1979
to
1982
and
from
1989
to'1991.
The
first
period
is
related
to
the
thesis
work
of
Blom
(1982),
Fottland
(1982)
and
Aarrestad
(1985).
In
the
second
period
new
analyses
were
done
in
order
to
increase
the
number
of
samples
and
include
new
areas,
especially
the
outer
and
inner
parts
of
the
Hardan-
gerfjord.
Of
the
162
releves
used
in
this
study,
Blom
made
43
releves
and
Fottland
made
36.
The
remaining
83
releves
were
made
by
the
author.
450
Nord.
J.
Bot.
20(4)
2000
Fig.
1.
Map
of
Hordaland
county
and
the
46
localities
investigated.
0
0
9
0
'
66
0
'0,
0
0
oo
A
OSterIly
Bergen
Os
Stord
ysnes
00
u
Sam
at
rr
r
Kvam
Kvinnherad
Granvin
.
•Ul
ik
0
10
20km
Selection
of
area
and
sites
The
area
has
been
selected
in
order
to
cover
the
main
variation
in
vegetational
composition
of
the
broad-
leaved
deciduous
forests
from
west
to
east
in
Hordaland
county.
The
westernmost
areas
with
small
stands
of
Corylus
avellana
(Stord,
Bonilo
and
islands
outside
Bergen)
have
been
avoided
becausC
of
strong
human
impact.
This
vegetation
type
has
been
investigated
previously
by
Romberg
&
Ovstedal
(1987)
and,
more
recently,
by
Smtersdal
&
Birks
(1993).
Within
the
selected
area
as
many
localities
as
possible
with
broad-leaved
deciduous
forests
have
been
visited,
except
localities
far
away
from
roads
because
of
the
difficulty
of
carrying
heavy
soil
sampling
equipment
for
environmental
studies.
To
be
analysed,
the
forests
must
support
stands
of
vegetation
where
the
dominant
tree
species
is
Corylus
avellana,
Fraxinus
excelsior,
Ulmus
glabra
or
Tilia
cordata,
or
a
combination
of
these
species.
Forests
with
heavy
and
recent
cultural
disturbance
were
avoided.
Vegetation
description
-
field
methods
The
vegetation
was
analysed
in
homogeneous
stands
based
on
the
homogeneity
-principle
of
Nordhagen
(1943)
and
Dahl
(1956).
Quadratic
or
rectangular
sample
plots
(releves)
were
subjectively
laid
out
in
the
stands.
Where
the
stands
were
large
enough,
areas
of
25
m
2
were
analysed.
No
releve
was
less
than
16
m
2
.
An
area
of
25
m
2
is
generally
considered
large
enough
to
cover
the
minimal
area
of
a
forest
community
(Fremstad
1979,
Kielland-Lund
1981).
In
general,
one
releve
was
analysed
in
each
stand.
A
few
larger
stands
were
analysed
with
more
than
one
releve,
however,
with
several
meters distance
between
the
releves.
The
collective
cover
of
each
layer
(tree,
shrub,
field
and
ground
layer)
was
estimated
as
a
percentage
of
the
sample
area,
and
species
abundances
were
recorded
according
to
the
Hult-Sernander
scale
(Du
Rietz
1921).
In
addition,
species
that
belonged
to
the
stand
but
occurred
just
outside
the
releve
were
given
a
cover
-
abundance
value
of
1.
This
concerns
very
few
species
Nord.
J.
Bot.
20(4)
2000
451
growing
one
or
two
meter
outside
the
releve
under
the
same
ecological
conditions
as
those
inside
the
releve.
No
such
individuals
have
been
recorded
in
connection
with
more
than
one
releve.
If
Anemone
nemorosa,
Gagea
lutea
or
Ranunculus
ficaria
were
observed
during
the
spring
aspect
but
not
when
the
analyses
were
done,
they
were
also
given
a
cover
-abundance
value
of
1.
In
general
one
should
not
mix
stand
level
and
releve
level
species
in
the
same
gradient
analysis,
since
such
analyses
should
be
done
at
one
scale
at
a
time
(e.
g.
Okland
1990).
For
optimal
numerical
analysis
these
species
could
then
have
been
removed
from
the
data
set.
However,
the
method
of
including
additional
species
was
used
in
the
plant
sociological
work
of
Blom
(1980,
1982)
in
order
to
include
species
that
probably
was
a
part
of
the
analysed
plant
community.
They
were
used
to
calculate
constancy
classes
of
species,
and
thus
they
are
a
part
of
the
phytosociological
classification
of
the
broad-leaved
deciduous
forests.
In
order
to
compare
the
earlier
plant
sociological
classification
with
the
numerical
classification
presented
in
this
work,
these
species
occurrences
have
therefore
been
included
in
the
data
set.
Data
treatment
The
vegetation
data
(162
releves
x
281
species)
with
Hult-Sernander
cover
abundances
were
arranged
in
a
sample
-by
-species
data
matrix
in
condensed
format
using
the
computer
program
ECOSURVEY
(Carleton
1986). Tree
and
shrub
species
that
occurred
in
more
then
one
vegetation
layer
were
summed
over
all
layers.
The
computer
program
TWINSPAN
version
2.1a
(Hill
1979a),
modified
by
ter
Braak
(1983)
and
H.
J.
B.
Birks,
was
used
to
classify
the
vegetational
data,
and
CANOCO
version
3.12
(ter
Braak 1988,
1990)
was
used
for
ordination
analysis.
The
ordination
diagrams
were
made
with
the
computer
program
CANODRAW
version
3.0
(Smilauer
1992).
Classification
of
vegetation
In
the
TWINSPAN
classification
five
pseudospecies
were
used
with
1,
2,
3,
4
and
5
as
cut
levels,
corresponding
to
the
Hult-Semander
cover
abundance
classes.
Allium
ursinum
was
given
a
non
-negative
weight
of
0.5,
due
to
big
seasonal
variation
in
abundance
of
this
species
within
the
releves
recorded
from
early
summer
to
early
autumn.
All
other
choices
in
the
TWINSPAN
settings
were
set
to
default
values.
In
the
vegetation
table,
based
on
this
classification,
average
cover
is
estimated
following
the
method
described
in
Sjers
(1954)
and
Malmer
(1962).
The
final
TWINSPAN
sample
groups
are
subjectively
chosen
due
to
the
strength
in
the
divisions
(eigenvalues)
and
evaluation
of
the
species
composition
between
the
groups.
The
differences
in
species
composition
were
also
evaluated
by
running
a
default
canonical
correspondence
analysis (CCA)
(ter
Braak
1986,
1987)
with
the
sample
-by
-species
data
matrix
as
the
species
data,
and
the
sample
classification
into
TWINSPAN
groups
(sample-by-TWINSPAN
groups
data
matrix)
as
the
"environmental"
data.
An
unrestricted
Monte
Carlo
permutation
test
with
99
random
permutations
was
used
to
investigate
whether
the
differences
in
species
composition
between
the
groups
were
statistically
significant
or
not.
Indirect
gradient
analyses
Correspondence
analysis
(CA)
(Hill
1973,
1974)
has
been
used
to
determine
the
major
patterns
of
variation
in
the
species
composition
and
to
interpret
this
variation
in
light
of
environmental
variables.
Due
to
the
uncertainty
in
recording
and
determining
of
rare
bryophyte
taxa,
caused
by
three
different
recorders
with
different
bryological
knowledge,
rare
species
were
downweighted.
Otherwise
default
options
were
chosen.
Unimodal
response
model
has
been
chosen
since
the
length
of
the
first
axis
in
a
detrended
correspondence
analysis
(DCA)
(Hill
1979b;
Hill
&
Gauch
1980)
of
the
total
data
was
above
2
standard
deviation
units,
as
recommended
by
Ter
Braak
&
Prentice
(1988).
CA
was
prefered
in
front
of
DCA
since
there
was
no
'arch
effect'
(Gauch
et
al.
1977)
or
'horseshoe
effect'
(Kendall
1971)
detected
in
CA,
implying
real
structures
on
the
two
first
axes
(Hill
&
Gauch
1980).
Nomenclature
-
species
and
plant
community
concepts
Taxonomic
nomenclature
follows
Lid
(1985)
for
vascular
plants,
Smith
(1978)
for
mosses
and
Grolle
(1976)
for
liverworts.
In
some
releves
it
was
impossible
to
distinguish
between
Epipactis
helleborine
and
E.
atrorubens.
These
two
species
have
therefore
been
lumped
together
in
the
analyses
and
named
Epipactis
spp.
However,
Epipactis
helleborine
is
the
most
common
species.
Quercus
spp.
includes
probably
both
Quercus
robur,
Q.
petraea
and
hybrids
of
these
species.
There
may
be
some
confusion
concerning
the
species
concepts
used
in
the
interpretation
of
TWINSPAN
results
compared
with
concepts
used
in
syntaxonomical
analyses.
In
TWINSPAN
context
preferential
species
are those
pseudospecies
and species
that
are
at
least
452
Nord.
J.
Bot.
20(4)
2000
BROAD-LEAVED
DECIDUOUS
HILL
FOREST
N=162
Cirriphyibm
poliferurri
I
Sifrnedioica
1
M=3
.0
13=3
FRAXINUS
-
CIRRIPHYliUM
PILIFERUM
FOREST
N=53
Cardamine
pratensis
1
Arer
preudopIaranus
1
Thelypteris
phegopteris
1
Athyrium
tiliK-temina
2
Mnium
h0mum
1
•00
FRAXINUS
ACER
-
ALNUS
ATNYNUM
TYPE
N=16
-0.191
t.thnus
glabra
1
B=7
•01
FRAXINUS
-
ULPAUS
TYPE
N=67
M.2
0.4
Iltathypodium
sylvaticurn
1
1
CORYL
US
311ACHYPODIUM
5
YLVATICUM
FOREST
N=79
litmus
giabra
1
Orchis
fnascula
1
Fdtpendula
ulmaria
1
k=0.21
Tilla
cordata
2
Thuidiurn
tamariscinum
1
hothecium
mynsurtudes
1
ht=1
M.
B=3
8=3
•10
•11
CORYLUS
-
ULMUS
-
ORCHID
TYPE
N.S1
110
N=20
CORYLUS
TILIA
FESTUCA
ALTISS1MA
LLIZULA
SYLVATiCA
TYPE
N
74
j.=0.270
Cardamine
praterrsis
1
•111
CORYLUS
-
POPULUS
G4
RDAM1NE
ARATENSiS
TYPE
N,4
Fig.
2.
TWINSPAN
classification
of
162
releves
into
sample
groups
(*)
and
plant
communities,
along
with
the
indicator
species,
eigenvalue
(X),
number
of
misclassified
(M)
and
borderline
(B)
samples
in
each
division.
twice
as
likely
to
occur
on
one
side
as
on
the
other
in
each
TWINSPAN
division,
and
they
occur
at
least
in
20
%
of
the
samples
on
the
preferential
side.
Indicator
species
are
the
most
highly
of
all
preferential
species
in
each
TWINSPAN
division.
A
differential
species
is
one
with
clear
ecological
preferences,
so
that
its
presence
can
be
used
to
identify
particular
environmental
conditions
(Hill
1979a).
They
can
be
derived
from
TWINSPAN
species
groups
and
used
to
separate
plant
communities
and
are
almost
absent
in
the
other
plant
communities
separated
by
TWINSPAN
calculations
(cf.
Odland
1991).
They
consist
both
of
indicator
species
and
strongly
preferential
species,
but
also
species
that
occur
in
less
than
20
%
of
the
samples
on
the
preferential
side.
Constant
species
occur
with
80
-
100
%
frequencies
in
the
vegetation
types
derived
by
the
TWINSPAN
classification.
Species
with
a
high
frequency
and
a
high
cover
in
a
vegetation
type
are
here
called
dominant
species,
while
characteristic
species
have
more
or
less
restricted
affinities
to
different
groups
of
samples.
They
may
include
both
indicator
species,
differential
species,
preferential
species,
constant
species
and
dominant
species.
The
plant
communities
derived
by
TWINSPAN
are
described
and
named
at
two
hierarchical
levels,
forest
as
the
highest
level
and
type
as
the
lowest
level.
They
mainly
represent
the
two
first
levels
in
the
TWINSPAN
division
and
are
named
from
the
dominant
tree
species,
indicator
species and
other
differential
species.
Results
Plant
communities
and
TWINSPAN
classification
Plant
communities
are
separated
in
each
TWINSPAN
division
at
different
hierarchical
levels
(Fig.
2),
and
five
sample
groups
(forest
types)
are
described
by
their
fl
oristic
composition
in
a
simplified
TWINSPAN
table
(Table
1).
The
species
are
listed
downwards
as
in
an
original
TWINSPAN
table,
but
with
the
frequency
and
average
cover
of
each
species
in
the
selected
TWINSPAN
sample
groups.
An
unrestricted
Monte
Carlo
permutation
test
of
the
axes
in
a
CCA
of
the
species
data
in
relation
to
the
TWINSPAN
-
sample
groups,
found
the
differences
in
species
composition
between
the
groups
in
each
TWINSPAN
division
to
be
statistically
significant
at
p
5.
0.01.
Nord.
J.
Bot.
20(4)
2000
453
Table
1.
Simplified
TWINSPAN
table.
Frequency
and
average
cover
of
species
in
the
forest
types
(TWINSPAN
sample
groups
*)
and
species
groups
(A
-L).
n
=
number
of
releves
in
each
sample
group.
Species
with
1,
2
or
3
occurrences
in
the
total
data
set,
except
species
with
a
frequency
more
than
10%
within
a
sample
group
and
a
few
species
(*)
referred
especially
in
the
text,
have
been
removed
from
the
table.
Tree
and
shrub
layer
species
in
bold
letters.
Species
Frax-Ace-
A
ln-Athy
*00
n
=
16
Frax-Ulm
*01
n
=
67
Cor-Ulm-
orchid
*10
n
=
51
Cor-Til-
Fest-Luz
*110
n
=
24
Cor-Pop-
Species
Card
groups
*111
n
=
4
Carex
muricata
*
-
3
1
Lapsana
communis
6
1
12
1
2
1
Actaea
spicata
24
1
Adoxa
moschatellina
*
3
1
Alchemilla
vulgaris
6
1
16
1
2
1
Chrysosplenium
oppositifolium
10
1
Gagea
lutea
16
1
Paris
quadrifolia
24
1
2
1
A
Chrysosplenium
alternifolium
19
1
36
1
Dryopteris
carthusiana
10
1
4
1
Impatiens
noli-tangere
15
2
Matteuccia
struthiopteris
46
4
2
1
Milium
effusum
13
1
4
1
Ranunculus
repens
6
1
10
1
Urtica
dioica
6
1
37
2
6
1
Poa
trivialis
25
1
43
1
25
1
Cardamine
fl
exuosa
25
1
36
1
6
1
Cirriphyllum
piliferum
94
1
91
1
10
1
4
1
Crepis
palludosa
13
1
21
1
4
1
Equisetum
pratense
19
3
16
3
Hokus
mollis
25
1
18
1
2
1
Brachythecium
rivulare
56
1
24
1
-
4
1
Plagiomnium
affine
44
1
30
1
2
1
4
1
B
Rhytidiadelphus
subpinnatus
19
1
13
1
-
Silene
dioica
100
2
63
2
14
1
25
1
Stellaria
nemorum
69
3
48
4
-
Galeopsis
bifida
56
1
27
1
2
1
8
1
Lophocolea
bidentata
63
1
27
1
4
1
Plagiothecium
denticulatum
44
1
15
1
2
1
4
1
Rhytidiadelphus
squarrosus
63
2
30
1
2
1
Acer
pseudoplatanus
81
2
10
2
2
1
4
1
Cardamine
pratense
81
2
18
1
100
1
Equisetum
sylvaticum
13
1
C
Mnium
spinosum
13
1
Ranunculus
acris
13
2
4
1
Brachytheciurn
rutabulum
56
1
42
1
16
1
13
1
50
1
Geum
rivale
63
2
39
1
4
1
100
2
Hylocomium
brevirostre
44
1
22
1
2
1
13
2
25
1
Hylocomium
umbratum
19
1
6
1
25
1
Plagiochila
asplenioides
81
1
40
1
4
1
8
1
75
1
Rhytidiadelphus
triquetrus
69
3
64
3
12
1
29
1
100
1
454
Nord.
J.
Bot.
20(4)
2000
Ribes
rubrum
6
1
4
1
25
1
Circaea
alpina
81
2
43
1
16
1
4
1
-
Filipendula
ulmaria
81
4
88
2
57
1
8
1
25
1
Oxalis
acetosella
100
3
93
2
39
1
29
2
100
2
Rhizomnium
punctatum
50
1
19
1
12
1
8
1
Alums
incana
75
4
45
3
20
1
8
1
Dryopteris
dilatata
38
1
6
1
8
1
Rumex
acetosa
38
1
1
1
2
1
-
Plagiomnium
undulatum
63
1
75
1
14
1
17
1
-
Ranunculus
auricomus
50
1
66
1
14
1
25
1
25
1
Ranunculus
fi
caria
19
1
31
2
12
1
-
-
Brachythecium
cf
oedipodium
4
1
4
1
Dentaria
bulbifera
25
1
-
21
1
-
Plagiochila
porelloides
22
1
2
1
17
1
-
Anomodon
rugelii
-
9
1
4
1
-
-
Stachys
sylvatica
38
2
73
2
43
2
8
1
-
Campanula
latifolia
40
1
18
2
4
1
Geranium
robertianum
31
1
51
1
22
1
13
1
-
Geum
urbanum
56
1
81
2
53
1
17
1
75
1
Alums
glutinosa
13
2
4
1
6
1
-
Athyrium
fi
lix-femina
100
4
87
2
88
2
83
1
100
2
Circaea
intermedia
69
1
36
1
24
1
4
1
Deschampsia
cespitosa
94
1
52
1
43
1
25
1
100
Agrostis
capillaris
19
1
3
1
6
1
8
1
Rubus
idaeus
100
2
48
1
37
1
38
1
25
1
Rhytidiadelphus
loreus
69
1
16
1
6
1
25
1
100
1
Salix
caprea
13
2
4
2
4
2
8
1
D
Thelypteris
phegopteris
100
2
45
1
29
1
46
1
100
3
Dicranum
majus
25
1
-
4
1
25
1
Hylocomium
splendens
31
2
6
1
2
1
21
1
100
1
Ranunculus
platanifolius
-
7
1
-
17
1
Eurhynchium
praelongum
88
1
91
1
43
1
38
1
100
1
Eurhynchium
striatum
75
2
84
2
59
1
79
1
75
1
Gymnocarpium
dryopteris
31
1
12
1
4
1
13
1
Plagiothecium
nemorale/succulentum
50
1
39
1
20
1
33
1
Thuidium
delicatulum
50
1
49
1
27
1
42
1
25
1
Thuidium
tamariscinum
100
2
79
1
37
1
96
1
100
2
Porella
cordaeana
-
12
1
2
1
17
1
Scrophularia
nodosa
6
1
10
1
6
1
50
1
Veronica
chamaedrys
25
1
43
1
16
1
42
1
Anemone
nemorsa
100
2
72
2
53
1
92
1
100
3
Dryopteris
filix-mas
38
2
55
1
35
1
75
1
25
1
Homalia
trichomanoides
6
1
48
1
27
1
42
1
Hypericum
maculatum
6
1
7
1
4
1
8
1
Thamnobryum
alopecurum
-
6
1
4
1
4
1
Anomodon
attenatus
16
1
16
1
4
1
Dactylis
glomerata
-
33
1
31
1
21
1
Fraxinus
excelsior
100
4
100
4
98
3
92
2
100
3
Geranium
sylvaticum
75
1
70
1
73
1
33
1
75
1
Pellia
epiphylla
19
1
13
1
12
1
8
1
25
1
Plagiomnium
cuspidatum
-
15
1
12
1
4
1
Poa
nemoralis
31
1
75
1
33
1
63
1
50
1
Polystichum
acuelatum
-
9
1
8
1
Ulmus
glabra
6
3
76
4
82
4
42
3
Valeriana
sambucifolia
69
2
52
1
59
1
38
1
25
3
Nord.
J.
Bot.
20(4)
2000
455
Angelica
sylvestris
19
1
37
1
24
1
21
1
Conopodium
majus
38
2
48
1
45
1
17
1
Epilobium
montanum
75
1
78
1
49
1
29
1
Prunus
padus
100
2
84
2
75
1
42
1
100
1
Brachythecium
glareosum
6
1
2
1
4
1
Plagiomnium
elatum
-
9
1
6
1
-
Polystichum
braunii
63
1
31
1
25
1
Plagiomnium
medium
7
1
6
1
-
E
Vicia
sepium
6
1
54
1
33
1
17
1
Conocephalum
conicum
6
1
4
1
Lysimachia
nemorum
13
1
8
1
Isopterygium
elegans
19
1
3
1
-
13
1
25
1
Mnium
hornum
69
1
15
1
4
1
29
1
75
1
Polytrichum
formosum
13
1
1
1
-
13
1
-
F
Luzula
pilosa
25
1
-
2
1
8
1
50
1
Bryum
spp.
6
1
3
1
2
1
13
1
Lophozia
spp.
6
1
3
1
2
1
8
1
Cystopteris
fragilis
6
1
10
1
4
1
-
Roegneria
canina
33
1
43
1
8
1
Carex
sylvatica
34
1
43
1
38
1
25
1
G
Circaea
lutetiana
6
4
4
1
17
3
Galium
odoratum
45
3
57
2
75
2
Vicia
sylvatica
4
1
4
1
4
1
Brachythecium
populeum
25
1
25
1
31
1
25
1
Festuca
gigantea
-
3
1
-
8
1
Atrichum
undulatum
19
1
34
1
47
1
54
1
50
1
Lophocolea
heterophylla
6
1
1
1
4
1
4
1
Sorbus
aucuparia
50
1
49
1
78
1
63
1
100
1
Fragaria
vesca
38
1
31
1
59
1
63
1
100
1
Deschampsia
fl
exuosa
56
1
3
1
10
1
50
1
50
1
Hypnum
cupresssiforme
1
1
2
1
13
1
Isothecium
myosuroides
13
1
9
1
14
1
50
1
100
2
Luzula
sylvatica
31
1
16
1
14
1
50
3
Populus
tremula
-
7
1
6
2
8
1
100
3
H
Betula
pubescens
25
1
9
1
14
2
33
3
75
2
Bryoerythrophyllum
recurvirostrum
1
1
2
1
8
1
Calypogeia
fissa
6
1
1
1
6
1
13
1
Fissidens
cristatus
18
1
37
1
58
1
Hypnum
spp.
3
1
6
1
17
1
Isothecium
myurum
6
1
16
1
51
1
92
1
75
1
Melica
nutans
13
1
25
1
50
1
100
1
Ctenidium
molluscum
39
1
63
1
83
1
100
2
Polygonatum
verticillatum
9
1
16
1
21
1
100
2
Convallaria
majalis
6
1
13
2
75
1
Rubus
saxatilis
6
1
6
1
22
1
21
2
100
2
Anthoxanthum
odoratum
13
1
Dicranum
scoparium
13
1
Festuca
altissima
15
1
43
1
83
4
Lathyrus
niger
*
4
1
I
Lathyrus
vernus
*
4
1
456
Nord.
J.
Bot.
20(4)
2000
Lonicera
periclymenum
6
1
-
12
1
42
1
50
1
Melampyrum
pratense
-
2
1
21
1
25
1
Tilia
cordata
10
2
31
3
96
5
Blechnum
spicant
1
1
4
1
13
1
-
Hieracium
spp.
6
1
8
1
8
1
75
1
Solidago
virgaurea
19
1
7
1
43
1
75
1
75
1
Veronica
officinalis
18
1
33
1
-
Hypericum
montanum
6
1
13
1
Neckera
complanata
4
1
8
1
Neottia
nidus-avis
*
-
4
1
4
1
J
Quercus
robur/petraea
16
2
29
1
Rosa
spp.
16
1
20
1
Fissidens
bryoides
1
1
16
1
17
1
Metzgeria
furcata
1
1
6
1
8
1
Tortella
tortuosa
1
1
12
1
8
1
-
Viburnum
opulus
7
1
63
1
58
1
75
2
Weissia
controversa
12
1
17
1
Allium
oleraceum
10
2
-
Bromus
ramosus
*
4
1
Campanula
rotundifolia
6
1
25
1
Campylium
stellatum
8
1
Carex
digitata
45
1
33
1
Carex
flacca
*
4
2
-
Dryopteris
pseudomas
10
1
8
1
Epipactis
helleborine/atrorubens
41
1
17
1
Fissidens
taxifolius
7
1
51
1
21
1
Hedera
helix
3
1
29
2
17
1
Heracleum
sphondylium
12
1
-
Ilex
aquifolium
25
1
13
1
K
Listera
ovata
6
1
27
1
8
2
Neckera
crispa
10
1
8
1
Primula
vulgaris
3
1
35
1
-
Rubus
nessensis
12
1
8
1
Sanicula
europaea
1
1
29
2
13
1
Brachythecium
sylvaticum
4
4
78
2
54
1
Bromus
benekenii
14
1
8
1
Eurhynchium
schleicheri
1
1
24
1
-
Homalothecium
sericeum
14
1
4
1
Taxus
baccata
35
1
17
1
Anomodon
viticulosus
3
1
14
1
Arctium
minus
1
1
8
1
Grimmia
hartmanii
1
1
6
1
Orchis
mascula
21
1
49
1
4
1
Succisa
pratensis
3
1
8
1
-
50
1
Taraxacum
spp.
16
1
51
1
8
1
25
1
L
Allium
ursinum
28
1
49
4
42
2
-
Corylus
avellana
19
1
58
4
100
5
100
4
100
5
Plagiothecium
cavifolium
4
1
12
1
4
1
Radula
complanata
3
1
10
1
8
1
Mycelis
muralis
13
2
19
1
35
1
50
1
Viola
riviniana
25
2
27
1
61
1
63
2
50
1
Nord.
J.
Bot.
20(4)
2000
457
Table
2.
Average
percentage
cover
of
tree,
shrub,
field
and
ground
layers
in
different
TWINSPAN
sample
groups
(plant
communities),
based
on
the
percentage
cover
of
each
layer
from
all
the
releves
in
each
TWINSPAN
group.
Sample
group
Plant
community
Tree
layer
Shrub
layer
Field
layer
Ground
layer
*0
Fraxinus-Cirriphyllum
piliferum
forest
65
11
78
51
*I
Corylus-Brachypodium
sylvaticum
forest
81
10
61
38
*00
Fraxinus-Acer-Alnus-Athyrium
type
59
22
76
35
*01
Fraxinus-Ulmus
type
66
9
78
54
*10
Corylus-Ulmus-orchid
type
82
9
64
55
*110
Corylus-Tilia-Festuca
altissima-Luzula
sylvatica
type
78
8
59
32
*
1 1 1
Corylus-Populus-Cardamine
pratensis
type
78
28
35
20
Fraxinus-Cirriphyllum
piliferum
forest
The
Fraxinus-Cirriphyllum
piliferum
forest
has
a
relatively
open
tree
layer
dominated
by
Fraxinus
excelsior,
a
dense
field
layer
of
tall
herbs
and
ferns
and
a
dense
bryophyte
layer
(Table
2).
Important
pre-
ferential
species
are
Cirriphyllum
piliferum,
Silene
dioica,
Alnus
incana,
Chrysosplenium
alternifolium,
Eurhynchium
striatum,
Plagiochila
asplenioides,
and
Stellaria
nemorum.
Species
in
TWINSPAN
species
groups
A
-
E
(Table
1)
have
strong
to
moderate
affinities
to
this
plant
community,
and
several
species
in
groups
A
-
C
can
be
regarded
as
differential
species
against
the
Corylus-Brachypodium
sylvaticum
forest.
Group
D
contains
species
that
occur
more
or
less
in
all
communities,
but
they
are
most
common
in
the
Fraxinus-Cirriphyllum
forest.
The
relative
open
tree
layer,
probably
caused
by
human
influence,
generates
a
high
number
of
heliophilous
species
(e.g.
Equisetum
pratense,
Holcus
mollis,
Cardamine
pratensis,
Geum
rivale,
Silene
dioica,
Rumex
acetosa,
and
Rhytiia-
delphus
squarrosus).
In
general,
the
species
com-
position
reflects
a
hygrophilous
and
slightly
thermophilous
plant
community.
Corylus-Brachypodium
sylvaticum
forest
The
Corylus-Brachypodium
sylvaticum
forest
has
a
dense
tree
layer
dominated
by
Corylus
avellana,
and
an
open
field
layer
characterised
by
a
large
proportion
of
small
herbs.
The
bryophyte
cover
is
relatively
low
(Table
2).
The
plant
community
is
characterised
by
its
indicator
species
Brachypodium
sylvaticum
and
by
additional
preferential
species
such
as
Allium
ursinum
(with
high
cover),
Carex
digitata,
Corylus
avellana,
Epipactis
spp.,
Festuca
altissima,
Ilex
aquifolium,
Taxus
baccata,
Tilia
cordata,
and
Veronica
officinalis.
Several
species
in
species
groups
I
-
K
can
be
regarded
as
differential
species
against
the
Fraxinus-
Cirriphyllum
piliferum
forest,
while
species
in
group
H
are
somewhat
more
frequent
in
this
plant
community.
The
dense
tree
layer
leads
to
lower
abundances
of
heliophilous
species and
a
larger
proportion
of
shade
tolerant
woodland
species,
such
as
Carex
digitata,
C.
sylvatica,
Festuca
altissima,
F.
gigantea,
Brachypo-
dium
sylvaticum,
Allium
ursinum,
Epipactis
spp.,
Sanicula
europaea,
and
Viola
riviniana.
Altogether
the
species
composition
indicates
a
more
thermophilous
and
less
hygrophilous
plant
community.
Fraxinus-Acer-Alnus-Athyrium
type
The
Fraxinus
-A
cer-A
lnus-Athyrium
type
has
a
relatively
open
tree
canopy
(average
cover
of
59%),
dominated
by
Fraxinus
excelsior
with
Acer
pseudo-
platanus
and
either
Alnus
incana
or
A.
glutinosa
as
constant
species.
The
average
cover
of
shrubs
is
relatively
high,
and
tall
ferns
of
Athyrium
filix-femina
dominate
the
field
layer.
It
is
characterised
by
several
indicator
species
(Fig.
2).,
Species
in
group
C
are
strongly
associated
with
this
type,
and
several
species
in
group
B
and
D
have
their
optimum
here;
mainly
less
nutrient
demanding
species.
Preferential
bryophytes
are
Brachythecium
rivulare,
Rhizomnium
punctatum,
Rhy-
tidiadelphus
loreus,
R.
squarrosus,
and
Lophocolea
bidentata.
Altogether
the
species
composition
reflects
a
slightly
thermophilous,
hygrophilous
and
mesotrophic
plant
community.
Fraxinus-Ulmus
type
The
tree
canopy
of
the
Fraxinus-Ulmus
type
is
dominated
by
Fraxinus
excelsior
and
Ulmus
glabra.
However,
Corylus
avellana
is
also
frequently
occurring.
458
Nord.
J.
Bogy.
20(4)
2000
It
is
somewhat
denser
than
the
canopy
of
the
Fraxinus-
Acer-Alnus-Athyrium
type,
but
still
relatively
open
with
an
average
cover
of
66%.
In
addition
to
its
indicator
species
it
is
characterised
by
additional
preferential
species
such
as
Actaea
spicata,
Allium
ursinum
(with
low
cover),
Campanula
latifolia,
Dactylis
glomerata,
Dentaria
bulbifera,
Galium
odoratum,
Homalia
trichomanoides,
Matteuccia
struthiopteris,
Orchis
mascula,
Paris
quadrifolia,
and
Polystichum
braunii,
all
rather
eutrophic
species.
Species
in
group
A
are
almost
exclusive
for
this
plant
community
(e.g.
Adoxa
moschatellina,
Chrysosplenium
oppositifolium,
Gagea
lutea,
and
Impatiens
noli-tangere).
Altogether
the
species
composition
reflects
a
slightly
thermophilous,
hygrophilous
and
eutrophic
plant
community.
Corylus-Ulmus-orchid
type
The
tree
canopy
of
the
Corylus-Ulmus-orchid
type
is
dominated
by
Corylus
avellana,
Ulmus
glabra
and
Fraxinus
excelsior,
the
two
firsts
with
the
highest
average
cover.
The
average
cover
of
the
tree
canopy
(82%)
is
the
highest
of
all
described
vegetation
types,
and
in
spring
and
early
summer
Allium
ursinum
forms
dense
stands
with
almost
no
other
species
in
the
field
layer,
except
Orchis
mascula
and
Ranunculus
ficaria.
Species
groups
J,
K
and
L
are
closely
associated
with
this
community,
which
also
has
a
high
number
of
species
with
low
to
medium
frequency.
The
thermophilous
and
highly
eutrophic
species
in
group
K
all
have
their
optimum
in
this
forest
type.
TWINSPAN
sample
group
*11
(Fig.
2)
is
separated
from
the
Corylus-Ulmus-orchid
type
by
its
indicator
species
Tilia
cordata,
Thuidium
tamariscinum
and
Isothecium
myosuroides.
However,
due
to
the
great
differences
in
species
composition
between
the
samples
within
this
sample
group,
the
next
division
has
been
chosen
to
describe
forest
type
communities.
Corylus-Tilia-Festuca
altissima-Luzula
sylvatica
type
The
Corylus-Tilia-Festuca
altissima-Luzula
sylvatica
type
has
generally
a
very
dense
tree
canopy
(78%)
dominated
by
Tilia
cordata,
with
constant
occurrence
of
Corylus
avellana
and
Fraxinus
excelsior,
however
with
somewhat
lower
abundance
values.
TWINSPAN
species
groups
I
and
J
contain
characteristic
species
for
this
community.
The
graminides
Festuca
altissima
and
Luzula
sylvatica
often
dominate
the
field
layer,
making
the
vegetation
type
somewhat
more
species
poor
than
the
Corylus-Ulmus-orchid
type.
Altogether
the
species
composition
reflects
a
thermophilous
and
medium
eutrophic
plant
community.
Corylus-Populus-Cardamine
pratensis
type
The
Corylus-Populus-Cardamine
pratensis
type
is
only
described
by
four
releves
from
one
locality
with
Cardamine
pratensis
as
an
indicator
species.
The
few
stands
analysed
have
a
tree
layer
totally
dominated
by
Corylus
avellana,
Populus
tremula
and
Fraxinus
excelsior,
with
Corylus
avellana
as
the
most
abundant
tree.
It
has
the
densest
shrub
layer
of
all
described
forest
types
(28%).
Species
associated
with
this
plant
community
are
found
most
frequently
in
species
groups
D,
F,
H
and
L.
Characteristic
species
are
Convallaria
majalis,
Polygonatum
verticillatum,
Rubus
saxatilis,
Sorbus
aucuparia,
Hylocomium
splendens
and
Isothecium
myosuroides.
Indirect
gradient
analysis
As
a
further
step
in
the
analysis
of
the
vegetational
data,
correspondence
analysis
(CA)
was
used
to
detect
gradients
in
the
species
composition,
and
to
evaluate
the
separation
of
the
TWINSPAN
groups,
and
their
consideration
as
plant
communities.
The
importance
of
the
axes,
shown
by
their
eigenvalues
(a),
is
relatively
high,
with
a
strong
first
axis
of
0.317.
The
next
three
axes
gradually
decrease
in
importance
with
eigenvalues
of
0.163,
0.130
and
0.103
respectively.
The
total
inertia
(T)
is
2.77,
and
the
percentage
of
variance
of
the
species
data
explained
by
the
axes
(X
/
T,
*
100)
are
quite
low
(11.5%,
5.8%,
4.7%
and
3.7%
respectively).
This
is
partly
due
to
large
amounts
of
noise
in
the
species
data,
and
is
very
common
for
abundance
data
in
general
(ter
Braak
1990).
However,
according
to
Okland
(1999)
the
total
inertia
is
probably
not
a
good
measure
of
the
total
variation,
and
the
amount
of
explained
variation
is
underestimated
by
using
the
ratio
of
the
eigenvalue
and
T.
The
first
CA
axis
separates
the
Cotylus-
Brachypodium
sylvaticum
forest
from
the
Fraxinus-
Cirriphyllum
piliferum
forest
with
almost
no
overlap
of
the
samples
(Fig.
3).
It
is
interpreted
as
a
climatic/soil
humidity
gradient
with
high
scores
of
thermophilous
and
dry
habitat
species
such
as
Hypericum
montanum,
Lathyrus
niger,
L.
vernus,
Bromus
ramosus,
Neckera
complanata,
N.
crispa,
Carex
flacca,
Neottia
nidus-
avis,
Rosa
spp.
and
Quercus
spp.,
all
related
to
releves
from
the
Corylus-Brachypodium
sylvaticum
forest.
Species
with
low
scores
on
CA
axis
one
are
Mnium
spinosum,
Equisetum
sylvaticum,
E.
pratense,
Chrysosplenium
alternifolium,
Stellaria
nemorum
and
Brachythecium
rivulare,
all
less
thermophilous
and
rather
hygrophilous
species
related
to
releves
from
the
Fraxinus-Cirriphyllum
piliferum
forest.
The
distribution
of
samples
on
the
second
CA
axis
Nord.
J.
Bot.
20(4)
2000
459
0
0
I
5
098
0
0
0
0
a
o
o
di
I:1
°
a
1.4
crab
D
o,
o
l
g
13
0
4
1
13
.
i
s
a
e
a
c
p.
a
a
a
.P3:
1
,
°
0
Ye
O.
4
*
II
10
%
di
w
-
Fig.
3.
Correspondence
analysis
(CA)
diagram,
axes
1
and
2,
of
the
sample
plots
with
their
TWINSPAN
sample
group
membership.
CA
axis
I
1.0
TWINSPAN
sample
groups
-
plant
communities
TWI
*00
0
Fraxinus-Acer-Alnus-Athyrium
type
TWI
*01
CI
Fraxinus-Ulmus
type
TWI
*10
Corylus-Ulmus-orchid
type
TWI
*110
Corylus-Tilia-Festuca
altissima-Luzula
sylvatica
type
TWI
*111
Corylus-Populus-Cardamine
pratensis
type
reflects
the
separation
of
plant
communities
in
the
second
TWINSPAN
division.
There
is
some
overlap
between
the
Fraxinus-Acer-Alnus-Athyrium
type
and
the
Fraxinus-Ulmus
type,
and
between
the
Corylus-
Tilia-Festuca
altissima-Luzula
sylvatica
type
and
the
Corylus-Uhnus-orchid
type,
but
still
they
represent
fairly
well
separated
units.
The
second
CA
axis
is
interpreted
as
a
soil
-richness
gradient
with
high
scores
of
less
nutrient
-demanding
species
such
as
Anthox-
anthum
odoratum,
Deschampsia
jlexuosa,
Dicranum
majus,
D.
scoparium
and
Hylocomium
splendens.
Species
with
low
scores
are
Allium
oleraceum,
Campylium
stellatum,
Carex
jlacca,
Listera
ovata,
Primula
vulgaris,
Orchis
mascula
and
Sanicula
europaea,
all
rather
nutrient
-demanding
species.
Further
TWINSPAN
divisions
were
added
into
the
+1.0
ordination
space,
but
these
sample
groups
gave
a
high
degree
of
overlap,
and
the
ordination
thus
supports
the
choice
of
the
final
TWINSPAN
groups.
Discussion
The
differentiation
of
plant
communities
by
TWINSPAN
The
first
TWINSPAN
division
of
the
162
releves
is
very
stable.
Different
choices
in
TWINSPAN
settings,
like
pseudospecies
cut
-levels,
weighting
of
species
and
weighting
of
samples
did
not
alter
the
sample
division.
The
relative
homogeneity
of
the
species
composition
within
the
TWINSPAN
sample
groups
is
reflected
in
460
Nord.
3.
Bot.
20(4)
2000
the
indirect
CA
ordination
by
a
clear
separation
of
the
samples
(Fig.
3).
However,
the
divisions
into
forest
types
are
somewhat
unstable.
Some
rearranging
of
intermediate
samples
occurred
with
different
choices
in
the
TWINSPAN
settings,
but
the
main
structure
remained
intact.
The
four
releves
making
up
the
Corylus-Populus-Cardamine
pratensis
type
are
very
different
from
the
other
samples
in
the
Corylus-
Brachypodium
sylvaticum
forest
and
thus
could
have
been
removed
from
the
analyses.
This
was
tried,
but
the
remaining
releves
were
classified
almost
identically
to
the
already
described
TWINSPAN
divisions.
Of
the
five
identified
vegetation
types,
the
Fraxinus-
Acer-Alnus-Athyrium
type,
the
Corylus-Tilia-Festuca
altissima-Luzula
sylvatica
type
and
partly
the
Corylus-
Populus-Cardamine
pratensis
type
fulfil
the
homo-
geneity
criteria
of
Raunkimr
(1934),
Nordhagen
(1943)
and
Dahl
(1956)
in
relation
to
the
«validity»
of
plant
communities.
In
the
Fraxinus-Ulmus
type
and
the
Corylus-Ulmus-orchid
type,
however,
the
constancy
class
V
(81-100%
frequency)
is
too
weak.
The
low
homogeneity
could
be
a
result
of
the
size
of
the
releves,
being
smaller
than
the
actual
minimum
area
of
these
vegetation
types.
Another
explanation
could
be
that
several
species
not
related
to
forest
communities
are
present
in
some
of
the
releves
due
to
stony
topography
and
cultural
impacts.
In
addition,
the
releves
are
sampled
in
an
area
with
a
broad
range
in
climate
and
geology,
and
the
number
of
species
with
low
and
medium
frequencies will
thus
be
high.
A
division
based
on
the
total
species,
as
in
the
TWINSPAN
classifi-
cation,
may
then
lead
to
a
lower
homogeneity
of
the
identified
vegetation
types.
Comparison
with
forest
communities
in
Scandinavia
and
in
Great
Britain
Most
of
the
releves,
classified
as
the
Fraxinus-
Cirriphyllum
piliferum
forest,
can
be
well
assigned
to
the
"alder
-ash
woodland"
in
the
Norwegian
classification
system
of
vegetation
types,
described
by
Fremstad
(1997).
They
belong
to
the
west
Norwegian
subtype,
where
Alnus
incana
is
more
prominent
than
Alnus
glutinosa
in
the
middle
and
inner
parts
of
the
fjords.
Most
of
the
releves
from
the
Corylus-
Brachypodium
forest
belong
to
the
western
subtype
of
the
"wych
elm
-
small
-leaved
lime
woodland"
(the
Ulmus
glabra-Tilia
cordata
woodland).
However,
Corylus
avellana
is
more
prominent
in
the
tree
layer
than
is
described
by
Fremstad
(op.
cit.).
In
the
inner
parts
of
the
fjords,
several
of
the
characteristic
species
of
the
western
subtype
disappear
or
become
less
frequent,
e.g.
Hedera
helix,
Ilex
aquifolium,
Cono-
podium
majus
and
Primula
vulgaris.
These
forests,
with
a
more
continental
character,
are
in
Fremstad
(1997)
assigned
to
the
eastern
subtype
of
the
"wych
elm
-
small
-leaved
lime
woodland".
The
forests
in
Hordaland
differ
from
those
in
Central
Norway
by
a
higher
frequency
of
thermophilous
species,
such
as
Bromus
benekenii,
B.
ramosus,
Festuca
altissima,
F.
gigantea,
Sanicula
europaea,
and
Allium
ursinum.
Tilia
cordata
is
almost
absent
in
Central
Norway,
and
Fraxinus
excelsior
is
less
dominant
than
in
the
west
Norwegian
forests
(cf.
Aune
1973;
Holten
1977, 1978,
1987;
Fremstad
1979).
Diekmann
(1994)
classified
the
Boreo-nemoral
deciduous
forests
of
Scandinavia
into
four
main
communities:
oligotrophic
oak
-forests,
mesotrophic
mixed
deciduous
forests,
eutrophic
elm-ash
forests,
and
eutrophic
alder
-ash
forests.
However,
releves
from
the
forests
investigated
in
Hordaland
were
not
available
in
his
numerical
classification,
and
the
described
communities
generally
reflect
the
vegetation
in
south-
east
Norway
and
southern
Sweden
with
a
lower
frequency
of
oceanic-
and
sub
-oceanic
species.
The
main
difference
between
the
west
Norwegian
communities
and
those
described
by
Diekmann
is
the
lower
frequency
or
a
lack
of
species
with
a
more
southern
or
south-eastern
distribution
pattern
(Fmgri
&
Danielsen
1996)
in
the
west
Norwegian
types,
e.g.
Acer
platanoides,
Anemone
ranunculoides,
Hepatica
nobilis,
Mercuralis
perennis,
and
Primula
veris.
Diekmann
(op.
cit.)
treated
the
forests
in
Hordaland,
referring
to
Blom
(1980)
and
Fottland
(1980),
as
"eutrophic
alder
-ash
forests"
and
"mesotrophic
mixed
deciduous
forests".
In
terms
of
species
composition,
the
Fraxinus-
Cirriphyllum
forest
(especially
the
Fraxinus-Ulmus
type)
could
well
be
assigned
to
the
"eutrophic
alder
-ash
forests"
on
the
basis
of
the
dominance
of
identical
woody
species
and
a
strong
resemblance
in
the
species
composition
of
tall
ferns
and
herbs.
On
the
other
hand,
the
Corylus-Brachypodium
forest
is
not
so
easily
fitted
into
Diekmann's
system.
Both
the
Corylys-Ulmus-
orchid
type
and
the
Corylus-Tilia-Festuca
altissima-
Luzula
sylvatica
type
show
some
affinities
to
Diekmann's
"mesotrophic
mixed
deciduous
forests",
due
to
the
presence
of
several
tree
species
in
the
stands
and
a
number
of
common
thermophilous
species.
However,
the
total
dominance
of
Allium
ursinum
in
many
of
the
west
Norwegian
stands,
especially
in
the
Corylus-Ulmus-orchid
type,
and
the
occurrence
of
Stachys
sylvatica
and
Campanula
latifolia,
which
in
Sweden
are
totally
absent
in
the
"mixed
deciduous
forests",
clearly
show
its
affinity
also
to
Diekmann's
"eutrophic
elm-ash
forests".
Due
to
the
oceanic
climate
of
the
investigated
area,
a
comparison
with
the
British
forest
communities
(Rodwell
1991)
is
appropriate,
even
though
these
communities
in
general
are
more
thermophilous
and
Nord.
J.
Rot.
20(4)
2000
461
Table
3.
Broad-leaved
deciduous
forest
communities,
their
syntaxonomical
equivalents
(main
association
and
subassociations
Community
Main
association
and
subassociation
represented
Fraxinus-Cirriphyllum
piliferum
forest
Fraxinus-Acer-Alnus-Athyrium
type
Fraxinus-Ulmus
type
Corylus-Brachypodium
sylvaticum
forest
Corylus-Ulmus-orchid
type
Corylus-Tilia-Festuca
altissima-
Luzula
sylvatica
type
Corylus-Populus-Cardamine
pratensis
type
Eurhynchio
striati-Fraxinetum
(Blom
1982)
Ovstedal
1985
E.
-F.
athyrietosum
Holcus-Alnus-Athyrium
facies
E.
-F.
athyrietosum
E.
-E
dentarietosum
bulbiferae
P.
-U.
circaetosum
lutetianae
Primulo
vulgaris-Ulmetum
glabrae
(Blom
1982)
Ovstedal
1985
P.
-U.
typicum
P.
-U.
allietosum
ursini
P.
-U.
polystichetosum
braunii
P.
-U.
festucetosum
altissimae
Ulmo-Tilietum
?
K.
-Lund
in
Seibert
1969
Melico-Quercetum
?
Bjernstad
1971
Melico-Betuletum
coryletosum
?
Aune
1973
=
not
approved
syntaxonomical
units,
more
affected
by
silviculture
than
those
on
the
west
coast
of
Norway.
Nevertheless,
there
is
a
striking
resemblance
in
species
composition
with
the
British
plant
communities,
especially
those
from
the
western
and
north-western
parts
of
Great
Britain
(Birks
1973;
Birse
1982,
1984).
The
Fraxinus-Cirriphyllum
forest
has
a
strong
affinity
to
the
more
humid
parts
of
the
British
"Fraxinus
excelsior-Sorbus
aucuparia-
Mercuralis
perennis
woodland"
(e.g.
the
"Crepis
paludosa
sub
-community").
Similarities
are
found
in
all
vegetation
layers,
especially
in
the
bryophyte
fl
ora
with
a
high
percentage
cover
of
species
such
as
Eurhynchium
praelongum,
E.
striatum
and
Cirriphyllum
piliferum.
The
Corylus-Brachypodium
forest
shows
some
similarities
in
species
composition
with
the
typical
sub
-
community
of
the
"Fraxinus
excelsior-Sorbus
aucu-
paria-Mercuralis
perennis
woodland"
on
somewhat
dryer
soils,
characterised
by
Brachypodium
sylvaticum
and,
more
occasionally,
Bromus
ramosus
and
Festuca
gigantea.
However,
the
affinity
is
even
greater
to
the
?
=
uncertain
syntaxonomical
equivalents
.
British
"Fraxinus
excelsior-Acer
campestre-
Mercuralis
perennis
woodland",
where
Corylus
avellana
is
the
most
frequent
woody
species
and
where
Taxus
baccata
and
Tilia
cordata
occur
occasionally.
Plant
communities
and
their
relation
to
syntaxonomy
As
emphasised
by
Diekmann
(1994)
the
Scandinavian
communities
are
difficult
to
fit
into
the
Braun-Blanquet
central
European
phytosociological
system,
because
the
northern
vegetation
types
were
not
taken
into
consideration
when
this
system
was
first
created.
An
overview
of
the
syntaxonomy
of
Norwegian
forest
vegetation
has
been
made
by
Kielland-Lund
(1994).
In
relating
the
plant
communities
created
by
TWINSPAN
classification
to
phytosociological
units,
I
have
used
the
syntaxonomy
of
Kielland-Lund
(op.
cit.),
based
on
Miller
(in
Oberdorfer
1992)
who
united
the
Quercetea
462
Nord.
J.
Bot.
20(4)
2000
represented)
and
their
placement
in
classes,
orders,
alliances
and
suballiances.
Alliance
and
suballiance
Order
Class
Alno-Ulmion
Br.
-B1.
et
Tx.
1943
Alnenion
glutinoso-incanae
Oberd.
1953
Tilio
platyphylli-Acerion
pseudoplatani
Klika
1955
Ulmo
glabrae-Tilienion
cordatae
K.
-Lund
1994
Quercion
robori-petraeae
Br.
-B1.
1932
Linnaeo-Piceion
Br.
-B1.
et
Siss.1939
corr.
Oberd.
1979
Fagetalia
sylvaticae
Pawl.
1928
Quercetalia
robori-petraeae
Tx.
(1931)
1937
em.
Th.
Mull.
in
Oberd.
1992
Querco-Fagetea
Br.
-B1.
et
Vlieg.
1937
em.
Oberd.
et
Th.
Mull.
1979
Vaccinio-Piceetalia
Br.
-B1.
1939
Vaccinio-Piceetea
em.
K.
-Lund
1967
Br.
-B1.
1939
robori-petraea
Br.
-B1.
1932
and
the
Querco-Fagetea
Br.-
Bl.
et
Vlieg.
1937
em.
Klika
1939.
All
the
described
forest
communities
can
be
assigned
to
the
mixed
deciduous
forest
class
Querco-Fagetea
(similar
to
Fraxino-Fagetea
Moor
(1976)
1978).
With
the
exception
of
the
four
releves
of
the
Corylus-
Populus-Cardamine
pratensis
type,
they
all
belong
to
the
order
Fagetalia
sylvaticae
Pawl.
1928
(Table
3).
This
is
due
to
the
presence
of
numerous
character
species
of
the
Fagetalia
(cf.
Ellenberg
1986).
Most
of
the
releves
from
the
Fraxinus-Cirriphyllum
piliferum
forest
belong
to
the
alliance
Alno-Ulmion
Br.B1.
et
Tx.
1943
(similar
to
Alno-Padion
Knapp
1942
em.
Mat.
et
Bor
1957)
and
the
suballiance
Alnenion
glutinoso-incanae
Oberd.
1953.
This
is
due
to
the
presence
of
several
of
the
alliance's
character
species,
such
as
Alnus
incana,
Matteuccia
struthiopteris,
Prunus
padus,
and
Stellaria
nemoreum.
The
west
Norwegian
association
Eurhynchio-Fraxinetum
(Blom
1982)
Ovstedal
1985
with
a
high
content
of
oceanic
species
(cf.
Fwgri
1960)
is
well
represented
in
this
group.
Some
releves
with
few
thermophilous
species and
dominance
of
Alnus
glutinosa
and
Prunus
padus
may
resemble
Alno-Prunetum
K.
-Lund
1971.
However,
they
differ
from
Alno-Prunetum
by
a
high
abundance
of
Fraxinus
exelsior
and
Ulmus
glabra.
A
few
releves
from
Granvin
and
Ulvik
in
the
inner
Hardangerfjord,
classified
to
the
Fraxinus-Cirriphyllum
piliferum
forest,
shows
low
frequencies
of
oceanic
species and
contains
eastern
species
such
as
Carex
muricata
and
Adoxa
moschatellina.
Thus
they may
belong
to
the
south-
eastern
association
Alno-Fraxinetum
K.
-Lund
1971,
which
according
to
Kielland-Lund
(1974)
is
present
in
the
area.
However,
occurrences
of
oceanic
species
such
as
Circaea
intermedia
and
Eurhynchium
striatum
and
the
lack
of
several
characteristic
species
of
the
south-
eastern
association
indicates
that
these
releves
are
not
a
part
of
the
Alno-Fraxinetum
association.
The
change
in
the
Fagetalia
forests
from
a
boreo-nemoral
character
in
the
western
and
middle
parts
of
the
fjords
to
forests
Nord.
.I.
Sot.
20(4)
2000
463
with
a
more
south
-boreal
character
in
the
innermost
areas
is
also
described
by
Austad
et
al.
(1985),
Austad
&
Skogen
(1990),
Swtersdal
&
Birks
(1993),
and
by
Fremstad
(1997).
The
Fraxinus-Acer-Alnus-Athyrium
type
can
be
compared
with
the
Holcus-Athyrium-Alnus
facies
of
the
poor
subassociation
Eurhynchio-Fraxinetum
athyrie-
tosum
described
by
Blom
(1982)
as
a
regenerating
stage
on
very
old
grazing
and
mowing
fields
(Table
3).
The
Fraxinus-Ulmus
type
consists
of
releves
that
belong
both
to
Eurhynchio-Fraxinetum
athyrietosum,
the
richer
Eurhynchio-Fraxinetum
dentarietosum
and
Primulo-
Ulmetum
circaetosum
lutetianae.
This
coincides
with
Kielland-Lund's
(1994)
proposal
of
including
Primulo-
Ulmetum
circaetosum
lutetianae
in
the
Eurhynchio-
Fraxinetum,
based
on
the
occurrence
of
several
hygrophilous
species.
Most
of
the
releves
in
the
Corylus-Brachypodium
sylvaticum
forest
belong
to
the
alliance
Tilio
platyphylli-Acerion
pseudoplatani
Klika
1955
(similar
to
Fagion-sylvaticae
(Luquet
1976)
Tx.
et
Diem
1936)
and
the
suballiance
Ulmo
glabrae-Tilienion
cordatae
K.
-Lund
1994.
Several
regional
character
species
of
the
alliance
(Vevle
1986;
Diekmann
1994)
are
present
in
the
releves,
such
as
Tilia
cordata,
Taxus
baccata,
Bromus
benekenii,
Neottia
nidus-avis,
Hedera
helix,
Festuca
altissima,
and
Sanicula
europaea.
The
oceanic
association
Primulo
vulgaris-Ulmetum
glabrae
(Blom
1982)
Ovstedal
1985,
occurring
in
the
outer
and
middle
parts
of
the
Hardangerfjord,
is
well
represented
in
this
group.
Some
releves
in
the
inner
Hardangerfjord
may
belong
to
the
eastern
Ulmo
glabrae-Tilietum
cordatae
K.
-Lund
in
Seibert
1969
due
to
the
low
abundance
of
oceanic
species.
However,
important
species
of
this
association
such
as
Viola
mirabilis,
Acer
platanoides,
Prunus
avium
and
Hepatica
nobilis,
are
not
present
in
the
analysed
releves.
The
Corylus-Ulmus-orchid
type
consists
of
releves
classified
by
Blom
(1982)
and
Fottland
(1982)
in
the
rich
subassociations
Primulo-Ulmetum
typicum,
Primulo-Ulmetum
allietosum
and
Primulo-Ulmetum
polystichetosum,
while
releves
from
the
subassociation
Primulo-Ulmetum
festucetosum
altissimae
are
placed
by
TWINSPAN
in
the
Corylus-Tilia-Festuca
altissima-
Luzula
sylvatica
type
together
with
possible
Ulmo-
Tilietum
stands
from
inner
Hardanger.
The
Corylus-Populus-Cardamine
pratensis
type
may
belong
to
Melico-Quercetum
Bjornstad
1971
in
the
relatively
poor
oak
-forest
alliance
Quercion
robori-
petraeae
Br.
-B1.
1932.
Even
if
oak
is
absent,
the
releves
contain
several
characteristic
species
of
the
oak
-forest,
such
as
Poa
nemoralis,
Melica
nutans,
Luzula
pilosa,
Deschampsia
flexuosa,
Viola
riviniana,
and
Lonicera
periclymenum.
Another
possible
equivalent
is
Melico-
Betuletum
coryletosum
Aune
1973,
a
west
-Norwegian
subassociation
in
Vaccinio-Picetea
Br.
-B1.
1939,
due
to
the
presence
of
Corylus
avellana,
Fragaria
vesca,
Viola
riviniana,
Hylocomium
splendens
and
Rhytidiadelphus
triquetrus.
Ovstedal
&
Rosberg
(1987)
related
several
hazel
coppices
on
the
western
coast
of
Norway
to
this
sub
-association,
and
the
analysed
stand
can
be
a
regenerating
stage
following
termination
of
human
impact
of
such
a
vegetation
type.
Conclusions
The
CA
analysis
clearly
separates
the
samples
from
the
first
TWINSPAN
division
along
the
first
ordination
axis,
and
the
Fraxinus-Cirriphyllum
piliferum
forest
and
the
Corylus-Brachypodium
sylvaticum
forest
can
thus
be
considered
as
two
distinct
different
plant
communities.
Both
classification
and
ordination
methods
can
justify
the
further
division
into
forest
types,
but
there is
clearly
a
gradient
structure
in
the
species
composition
between
these
plant
communities.
The
investigated
broad-leaved
deciduous
forests
show
many
similarities
in
species
composition
with
forest
communities
in
other
parts
of
Scandinavia
and
in
Great
Britain.
However,
they
are
distinct
different
from
these
by
a
higher
element
of
oceanic
species
and
less
abundance
of
thermophilous
species
more
widely
distributed
in
southern
parts
of
Europe.
In
relation
to
syntaxonomy
the
first
TWINSPAN
division
supports
Blom's
(1982)
main
separation
of
the
oceanic, broad-leaved
deciduous
hill
forests
into
a
hygrophilous
plant
community,
the
Eurhynchio-
Fraxinetum,
and
a
drier
and
more
thermophilous
community,
the
Primulo-Ulmetum.
On
the
other
hand,
further
division
of
the
TWINSPAN
produces
communities
that
do
not
support
the
proposed
sub
-
associations
of
the
two
communities.
The
species
composition
of
the
releves
from
the
inner
parts
of
the
Hardangerfjord
is
not
typical
for
the
south-eastern
Alno-Fraxinetum
and
Ulmo-Tilietum
associations,
which,
according
to
Kielland-Lund
(1994),
is
present
in
the
analysed
area.
I
believe
that
these
releves
reflect
a
transitional
stage
between
the
western
and
the
south-
eastern
associations.
Even
though
plant
communities
can
be
separated,
the
vegetation
in
the
west
Norwegian
boreo-nemoral
forests
is
reflected
by
gradients
in
species
composition,
both
within
plant
communities
and
between
communities.
Thus
the
described
plant
communities
must
not
be
considered
as
strictly
delimited
units,
but
more
as
a
means
of
describing
the
differences
in
the
vegetation
and
as
a
basis
for
understanding
the
underlying
environmental
gradients
influencing
the
species
composition
of
the
plant
communities.
464
Nord.
J.
Bat.
20(4)
2000
Acknowledgements
-
I
am
very
grateful
to
John
Birks,
Arnfinn
Skogen,
and
Arvid
Odland
for
advice
and
critical
reading
on
the
manuscript
and
to
Hillary
Birks
for
comments
on
the
English
text.
Two
reviewers
have
suggested
several
improvements,
and
I
would
also
like
to
thank
Hans
Blom
and
Mon
Fottland
for
the
permission
to
use
some
of
their
unpublished
vegetation
data
from
sites
where
I
have
collected
soil
samples.
Fieldwork
has
been
supported
by
the
Grolle
Olsen's
legat.
References
Aarrestad,
P.
A.
1985.
Samanheng
mellom
vegetasjon
og
jordsmonn
i
edellauvskogar
i
Bergensregionen
og
midtre
Hardanger.
-
Cand.
real.
thesis,
Univ.
Bergen.
Askvik,
H.
1976.
Hordalands
berggrunnsgeologi.
-
In:
Hartvedt,
G.
H.
(ed.),
Bygd
og
by
i
Norge,
Hordaland
og
Bergen.
Oslo,
pp.
100-110.
Aune,
B.
1993.
Temperaturnormaler,
normalperiode
1961-
1990.
-
Det
norske
meteorologiske
institutt,
Oslo.
Rapp.
02/93:
1-66.
Aune,
E.
1.
1973.
Forest
vegetation
in
Hemne
Ser-Trondelag,
Western
Central
Norway.
-
K.
norske
Vidensk.
Selsk.
Mus.
Misc.
12:
1-87.
Austad,
I.
&
Losvik,
M.
H.
1998.
Changes
in
species
composition
following
field
and
tree
layer
restoration
and
management
in
a
wooded
hay
meadow.
-
Nord.
J.
Bot.
18:
641-662.
-
&
Skogen,
A.
1988.
Havratunet
i
°stem)
,
kommune.
En
botanisk-okologisk
analyse
og
en
plan
for
istandsetting
og
skjetsel
av
kulturlandskapet.
-
Okoforsk
Rapp.
1988(13):
1-119.
-
&
Skogen,
A.
1990.
Restoration
of
a
deciduous
woodland
in
western
Norway
formerly
used
for
fodder
production:
effects
on
tree
canopy
and
fl
oristic
composition.
-
Vegetatio
88:
1-20.
-
,
Lea,
B.
0.
&
Skogen,
A.
1985.
Kulturpavirkete
edel-
lauvskoger.
Utproving
av
et
metodeopplegg
for
istandsetting
og
skjetsel.
-
Okoforsk
Rapp.
1985(1):
1-55.
Balle,
0.
1978.
Vegetasjonsokologiske
studier
i
vest-norske
lovskogslier,
med
swrlig
hensyn
pA
jordbrukets
innvirkning.
-
Cand.
real.
thesis,
Univ.
Bergen.
Birks,
H.
J.
B.
1973.
Past
and
Present
Vegetation
of
the
Isle
of
Skye:
a
PaleoecOlogical
Study.
-
Cambridge
University
Press,
London.
Birse,
E.
L.
1982.
The
main
types
of
woodlands
in
North
Scotland.
-
Phytocoenologia
10:
9-55.
-
1984.
The
Phytocoenonia
of
Scotland:
Additions
and
Revisions.
-
Macaulay
Institute
for
Soil
Research.
Aberdeen.
Blom,
H.
H.
1980.
Plantesosiologiske
undersokelser
av
edellauvskog og
beslektede
samfunn
pa
frisk
mark
i
Ytre
Hordaland.
-
K.
norske
Vidensk.
Selsk.
Rapp.
Bot.
Ser.
1980(5):
134-150.
-
1982.
Edellevsskogssamfunnene
i
Bergensregionen,
Vest
-
Norge.
-
Cand.
real.
thesis,
Univ.
Bergen.
Carleton,
T.
J.
1986.
ECOSURVEY
-
microcomputer
programs
for
the
analysis
of
ecological
survey
data
on
the
IBM-PC
and
compatibles.
-
Toronto.
Dahl,
E.
1956.
Rondane.
Mountain
vegetation
in
South
Norway
and
its
relation
to
the
environment.
-
Skr.
Norske
Vidensk.-Akad.
Mat.-Naturv.
K1.
1956(3):
1-374.
Oslo.
-,
Elven,
R.,
Moen,
A.
&
Skogen,
A.
1986.
Vege-
tasjonsregionkart
over
Norge
1:1
500
000.
-
Nasjonalatlas
for
Norge.
Statens
Kartverk.
Diekmann,
M.
1994.
Deciduous
forest
vegetation
in
Boreo-
nemoral
Scandinavia.
-
Acta
Phytogeogr.
Suec.
80:
1-112.
Du
Rietz,
G.E.
1921.
Zur
methodlogischen
Grundlage
der
Modernen
Pflanzensoziologie.
-
Akad.
Abh.
Uppsala.
Ellenberg,
H.
1986.
Vegetation
Mitteleuropas
mit
den
Alpen
in
okologischer
Sicht.
-
4.
Aufl.
Verlag
Eugen
Ulmer,
Stuttgart.
Fottland,
H.
1980.
Rik
lauvskog
i
midtre
Hardanger.
-
K.
norske
Vidensk.
Selsk.
Rapp.
Bot.
Ser.
1980(5):
127-133.
-
1982.
Edellauvskog
i
midtre
Hardanger.
-
Cand.
real.
thesis,
Univ.
Bergen.
Fremstad,
E.
1979.
Phytosociological
and
ecological
investigations
of
rich
deciduous
forests
in
Orkladalen,
Central
Norway.
-
Norw.
J.
Bot.
26:
111-140.
-
1983.
Role
of
Black
Alder
(Alnus
glutinosa)
in
vegetation
dynamics
in
West
Norway.
-
Nord.
J.
Bot.
3:
393-410.
-
1997.
Vegetasjonstyper
i
Norge.
-
NINA
Temahefte
12:
1-
279.
-
&
Elven,
R.
1996.
Fremmede
planter
i
Norge.
Platanlonn
(Acer
pseudoplatanus
L.).
-
Blyttia
2:
61-78.
Fwgri,
K.
1960.
Maps
of
distribution
of
Norwegian
vascular
plants.
I.
The
coast
plants.
-
Univ.
Bergen
Skr.
26:
1-134.
-
&
Danielsen,
A.
1996.
Maps
of
distribution
of
Norwegian
vascular
plants.
III.
The
southeastern
element.
-
Univ.
Bergen.
Fagbokforlaget
Bergen.
Forland,
E.
J.
1979.
Nedborens
hoydeavhengighet.
-
Klima
2:
3-24.
-
1993.
Nedbornormaler, normalperiode
1961-1990.
-
Det
norske
meteorologiske
institutt,
Oslo.
Rapp.
39/93:
1-63.
Gauch,
H.
G.,
Whittaker,
R.
H.
&
Wenthworth,
T.
R.
1977.
A
comparative
study
of
reciprocal
averaging
and
other
ordination
techniques.
-
J.
Ecol.
65:
157-174.
Grolle,
R.
1976.
Verzeichnis
der
Lebermoose
Europas
and
benachbarter
Gebiete.
-
Feddes. Repert.
87:
171-279.
Hill,
M.
0.
1973.
Reciprocal
averaging:
an
eigenvector
method
of
ordination.
-
J.
Ecol.
61:
237-249.
-
1974.
Correspondence
analysis:
a
neglected
multivariate
method.
-
Applied
Statistics
23:
340-354.
-
1979a.
TWINSPAN
-
a
FORTRAN
program
for
arranging
multivariate
data
in
an
ordered
two-way
table
by
classification
of
individuals
and
attributes.
-
Cornell
University,
Ithaca,
New
York.
-
1979b.
DECORANA
-
A
FORTRAN
program
for
detrended
correspondence
analysis
and
reciprocal
averaging.
-
Cornell
Univ.,
Ithaca,
New
York.
-
&
Gauch,
H.
G.
1980.
Detrended
correspondence
analysis:
an
improved
ordination
technique.
-
Vegetatio
42:
47-58.
Holten,
J.
I.
1977.
Floristiske
og
vegetasjonsokologiske
undersokelser
i
star-
og
nordeksponerte
tier
ved
Gjora
i
Sunndal.
-
Cand.real.
thesis,
Univ.
Trondheim.
-
1978.
Verneverdige
edellauvskoger
i
Trendelag.
-
K.
norske
Vidensk.
Selsk.
Mus.
Rapp.
Ser.
1978(4):
1-199.
-
1987.
Autecological
and
phytogeographical
investigations
along
a
coast
-inland
transect
at
Nordmore,
Central
Norway.
-
Dr.
philos.
thesis,
Univ.
Trondheim.
Kendall,
D.
G.
1971.
Seriation
from
abundance
matrices.
-
In:
Hodson,
F.
R.,
Kendall,
D.
G.
&
Tauto,
P.
(eds),
Mathematics
in
the
archaeological
and
historical
sciences.
Nord.
J.
Bot.
20(4)
2000
465
University
Press,
Edinburgh.
pp.
215-252.
Kielland-Lund,
J.
1974.
Skogsvegetasjon
i
Ulvikomradet.
-
Medd.
Geogr.
Inst.
Nor.
Handelsheysk.
Univ.
Bergen
32
B,
G.
-
1981.
Die
Waldgesellschaften
SO-Norwegens.
-
Phyt-
ocoenologia
9.
1/2:
53-250.
-
1994.
Syntaxonomy
of
Norwegian
forest
vegetation
1993.
-
Phytocoenologia
24:
299-310.
Korsmo,
H.
1974.
Ostfold,
Akershus,
Hedmark
og
Oppland.
Naturvernradets
landsplan
for
edellauvskogsreservater
i
Norge
I.
-
As,
NLH.
-
1975.
Hordaland,
Sogn
og
Fjordane
og
More
&
Romsdal.
Naturvernradets
landsplan
for
edellauvskogsreservater
i
Norge
4.
-
As,
NLH.
Kvamme,
M.
1988.
Pollenanalytiske
bidrag
til
edellovsskogenes
historie
i
Ser-Norge.
-
In:
Bretten,
S.
&
Relining,
0.
1.
(eds),
Fagmete
i
vegetasjonsokologi
pa
Kongsvoll
1988.
Univ.
Trondheim,
Vitensk.
mus.
Rapp.
Bot.
Ser.
1988-1,
pp.
19-26.
Lea,
B.
0.
1984.
Struktur
og
vegetasjonsdynamikk
i
en
rik
lauvskog
i
indre
Sogn.
-
Cand.
scient.
thesis,
Univ.
Bergen.
Lid,
J.
1985.
Norsk,
svensk,
finsk
fl
ora.
-
Det
norske
samlaget,
Oslo.
Losvik,
M.
1978.
Vegetasjonsklassifikasjon
og
kartlegging
med
sikte
pa
anvendelse
i
landskapsplanlegging
i
Bergensregionen.
-
Cand.
real.
thesis,
Univ.
Bergen.
Malmer,
N.
1962.
Studies
of
mire
vegetation
in
the
Archaean
Area
of
Southwestern
Gotaland
(South
Sweden).
I.
Vegetation
and
Habitat
conditions
on
the
Akhult
mire.
-
Opera
Bot.
7(1):
1-322.
Moen,
A.
1987.
The
regional
vegetation
of
Norway;
that
of
Central
Norway
in
particular.
-
Norsk
geogr.
Tidskr.
41:
179-226.
-
1998.
Nasjonalatlas
for
Norge:
Vegetasjon.
-
Statens
kartverk,
Henefoss.
Nordhagen,
R.
1943.
Sikilsdalen
og
Norges
fi
ellbeiter.
-
Bergens
Mus.
Skr.
22:
1-607.
Oberdorfer,
E.
1992.
Siiddeutche
Pflanzengesellschaften.
Teil
4:
Walder
and
Gebiische.
2.
Aufl.
-
Gustav
Fischer
Verlag,
Jena.
Odland,
A.
1991.
On
the
ecology
of
Thelypteris
limbosperma
-
a
synecological
investigation
of
T.
limbosperma-
dominated
stands
in
W
Norway.
-
Nord.
J.
Bot.
10:
637-
659.
-
1992.
A
synecological
investigation
of
Matteuccia
struthiopteris
-
dominated
stands
in
Western
Norway.
-
Vegetatio
102:
69-95.
Presch-Danielsen,
L.
1996.
Vegetation
history
and
human
impact
during
the
last
11
500
years
at
Lista,
the
southernmost
part
of
Norway.
Based
primarily
on
Professor
Ulf
Hafsten's
material
and
diary
from
1955-
1957.
-
Norsk
geogr.
Tidskr.
51:
85-99.
Raunkimr,
C.
1934.
The
life
form
of
plants
and
statistical
plant
geography.
-
Oxford
University
Press,
Oxford.
Rodwell,
J.
R.
(ed.)
1991.
British
Plant
Communities.
Vol.
1.
Woodlands
and
scrub.
-
Cambridge
University
Press,
Cambridge.
Rosberg,
I.
&
Ovstedal,
D.
0.
1987.
Phytosociology
and
soil
properties
of
Corylus
avellana
coppices
on
the
coast
of
western
Norway.
-
Nord.
J.
Bot.
7:
169-185.
Sjers,
H.
1954.
Slatteranger
i
Grangarde
Finnmark.
-
Acta
Phytogeogr.
Suec.
34:
1-184.
Smilauer,
P.
1992.
CanoDraw
User's
Guide,
version
3.0.
-
Microcomputer
Power,
Ithaca,
New
York,
USA.
Smith,
A.
J.
E.
1978.
The
moss
fl
ora
of
Britain
and
Ireland.
-
Cambridge.
Seetersdal,
M.
1994.
Rarity
and
species/area
relationships
of
vascular
plants
in
deciduous
woods,
western
Norway
-
applications
to
nature
reserve
selection.
-
Ecography
17:
23-38.
-
&
Birks,
H.
J.
B.
1993.
Assessing
the
representativeness
of
nature
reserves
using
multivariate
analysis:
Vascular
plants
and
breeding
birds
in
deciduous
forests,
western
Norway.
-
Biological
Conservation
65:
121-132,
-
,
Line,
J.
M.
&
Birks,
H.
J.
B.
1993.
How
to
maximize
biological
diversity
in
nature
reserve
selection:
Vascular
plants
and
breeding
birds
in
deciduous
woodlands,
western
Norway.
-
Biological
Conservation.
66:
131-138.
ter
Braak,
C.
J.
F.
1983.
Differential
weights
for
samples
and
attributes
in
the
ordination
program
DECORANA
and
the
classification
program
TWINSPAN.
-
Institute
TNO
for
mathematics,
Information
Processing
and
Statistics,
Report
C
82
ST
106
55:
1-10.
-
1986.
Canonical
correspondence
analysis:
a
new
eigenvector
technique
for
multivariate
direct
gradient
analysis.
-
Ecology
67:
1167-1179.
-
1987.
The
analysis
of
vegetation
-environment
relationships
by
canonical
correspondence
analysis.
-
Vegetatio
69:
67-77.
-
1988.
CANOCO
-
a
FORTRAN
program
for
canonical
community
ordination
by
(partial)
(detrended)
(canonical)
correspondence
analysis,
principal
components
analysis
and
redundancy
analysis
(version
2.1).
-
Technical
report
LWA-88-02.
Agricult.
Math.
Group,
Wageningen,
The
Netherlands.
-
1990.
Update
notes:
CANOCO
version
3.10.
-
Agricult.
Math.
Group,
Wageningen.
-
&
Prentice, I.C.
1988.
A
theory
of
gradient
analysis.
-
Adv.
ecol.
Res.
18:
271-317.
Ve,
S.
1940.
Skog
og
treslag
i
Indre
Sogn
fra
Lxrdal
til
Fillefjell.
Med
ei
utgreing
um
gran
i
Sogn.
-
Medd.
Vestl.
forst.
Forsoksst.
23:
1-224.
Ve,
S.
1941.
Bonden
buskapen
og
skogen
i
gamle
Vestlandsbygder.
-
Tidskr.
for
Skogbr.
49:
149-157
and
205-215.
Vevle,
0.
1986.
Norske
vegetasjonstypar.
-
Be,
Telemark.
Nistas
forlag.
Okland,
R.
H.
1990.
Vegetation
ecology:
theory,
methods
and
applications
with
reference
to
Fennoscandia.
-
Sommerfeltia
Suppl.
I
:
1-233.
-
1999.
On
the
variation
explained
by
ordination
and
constrained
ordination
axes.
-
J.
Veg.
Sci.
10:
131-136.
Ovstedal,
D.
0.
1985.
The
vegetation
of
Lindas
and
Austrheim,
western
Norway.
-
Phytocoenologia
13:
323-
449.
466
Nord.
J.
Bot.
20(4)
2000