Vegetatio 80: 167-181, 1989.
© 1989 Kluwer Academic Publishers. Printed in Belgium.
167
A classification of the deciduous forest of eastern North America
Carl D. Monk 1, Donald W. Imm 1, Robert L. Potter ~ & Geoffrey G. Parker 2
Botany Department and Institute of Ecology, University of Georgia, Athens, GA 30602, USA;
2 Smithsonian Environmental Research Center, P.O. Box 28, Edgewater, MD 21037, USA
Accepted 30.1.1989
Keywords: DCA, Indicator analysis, Ordination, TWlNSPAN
Abstract
Data from 300 forest stands, scattered over 29 states within the eastern North American deciduous forest,
were subjected to detrended correspondence analysis (DCA) and two-way indicator species analysis
(TWlNSPAN) in an effort to identify classifiable units. Most species are widespread which provide a
great deal of continuity in the vegetation.
The deciduous forest can be divided into three forest regions: (1) northern, (2) central and (3) southern.
The northern region corresponds to the hemlock-white pine-northern hardwood forest of Braun (1950).
The central region includes the beech-maple and oak-hickory forests. The beech-maple as identified here
includes the mixed mesophytic, beech-maple, maple-basswood and about half of the western mesophytic
forests of Braun (1950). The oak-hickory includes Braun's oak-hickory, oak-chestnut and about half of
the western mesophytic forests. The southern region coincides with the southern mixed hardwood forests.
Nomenclature: National list of scientific plant names. Soil Conserv. Service Techn. Bull. 159, 1982.
Introduction
The broad-leaved deciduous forests of eastern
North America occur within the region of typical
temperate, four-season climate (Walter 1985).
The deciduous forest reaches its northern limit in
central Minnesota and extends eastward to
southwest Quebec and from Maine to south central Florida and west into Louisiana and eastern
Texas. The northward distribution is restricted by
a short growing season ( < 120 days with mean
daily temperature > 10 °C) and low winter temperatures. Decreasing precipitation and longer
and more frequent droughts confine the forests'
westward extension to mesic slopes, ravines and
river margins. Forests at the southward limit are
restricted by warmer climate and particularly by
the absence of or a very short winter dormant
season.
Even though the basic growth of the dominant
tree species is broad-leaved deciduous, important
evergreen components are found throughout most
of the deciduous forest area, depending on habitat
and site history. Indeed, needle-leaved evergreens
are often dominants or co-dominants throughout
much of the deciduous forest in successional
situations. Broad-leaved evergreens become more
important in the southern region of the deciduous
forests as well as on nutrient-poor sites
throughout (Monk 1966; Monk et al. 1985).
Although no single tree species ranges
throughout the deciduous forest, 6 wide-ranging
168
e~
o
.=~ .= . ' ~
u,l ~.1
~
o
m
o
=
o
=:
>
.o
o
a
~
0
0
0
0
o
o
0
0
o
o
o
o
o
0
0
0
0
0
0
o
0
0
~o
~
•<
0
o
0
o
0
,'-.
r.
r.
o
o
o
o
o
o
o
N
N
~
0
o
o
0
~
e
o
e
o ~
~ N
"r,
=
e.
~
6
o
.o
~
0
<
,3
~
o
-
o , .-~
='~
~~.~
o ~"=
~
z
.~ z ~
"r=
~~ oo> ~~
~, o
-~if:
E
o
Z
Z
z
.~
~:
Z
0
.o
e~
...4
o
r~
o
if-
169
genera (Acer, Fagus, Fraxinus, Quercus Tilia and
Ulmus) are ubiquitous. The importance of Quercus
and Fagus led Clements (1916, 1928) and Weaver
& Clements (1938) to designate the deciduous
forest the Quercus-Fagus Formation.
The fioristic composition varies continuously
across the eastern deciduous forest and schemes
for its classification vary from author to author
(Table 1). The area that is recognized as the
eastern deciduous forest today was divided into 3
segments by Sargent (1884), Transeau (1905) and
Mattoon (1936). Clements (1916, 1928) and
Weaver & Clements (1938) considered the Lake
Forest as a separate unit from the deciduous
forest which they divided into maple-beech (Acer
-Fagus), oak-chestnut (Quercus-Castanea) and
oak-hickory (Quercus-Carya). The Lake Forest
Formation of Clements (1928) on Northern Pine
Belt of S argent (1884) centered in the Great Lakes
Region and occurring eastward into New York
and New England, is accepted by most modern
authorities as an integral part of the eastem
deciduous forest.
Mattoon (1936) spoke of the central
hardwoods, including Braun's (1950) mixed
mesophytic, western mesophytic, oak-hickory,
oak-chestnut, beech-maple and maple-basswood
(Acer-Tilia), 'northern region', basically the hemlock ( Tsuga)-white
pine
(Pinus)-northern
hardwoods and the 'southern region' where oakpine is transitional with the southeastem evergreen forest region. Oosting (1956), Polunin
(1960), Gleason & Cronquist (1964), Ktlchler
(1964) and Greller (1988) are in close agreement
with Braun's treatment, though some differences
exist. For example; Ktichler used the term
Appalachian oak instead of oak-chestnut and
Oosting expanded his oak-hickory to include the
coastal plain vegetation. Greller (1988) proposed
a high-elevation maple-beech-birch (Betula)buckeye (Aesculus) community type that he feels
is distinct from the mixed mesophytic forest.
Other treatments tend toward lumping. Barbour
et al. (1987) pool Braun's maple-basswood and
beech-maple into beech-maple and Braun's oakchestnut, oak-hickory and oak-pine into oakhickory. However, Barbour etal. (1987) still
recognized the southern mixed hardwoods of the
coastal plain. Daubenmire (1978) divided the
deciduous forest into theAcer saccharum Province
and the Quercus falcata Province. The mixed
mesophytic, western mesophytic, beech-maple,
maple-basswood and hemlock-white pine-northern hardwoods of Braun (1950) made up
Daubenmire's Acer saccharum Province, while the
oak-hickory, oak-chestnut, oak-pine and
southeastern evergreen forest regions made up his
Quercus falcata Province.
The deciduous forest of eastern North America
is composed of many stands dominated by relatively few tree species. Many studies have dealt in
varying detail with community types or vegetation
within certain geographical areas (i.e. Oosting
1942; Whittaker 1956; Curtis 1959; Robichand &
Buell 1973). Several authors have examined large
datasets from a specific region (Christensen 1988)
or community type (Ware 1982; Parker
1985,1987). No one has used ordination and
numerical classification techniques on stands distributed throughout the eastern deciduous forest
in an attempt to classify the forest vegetation in a
more objective way. That is the main objective of
this paper.
Study areas and Methods
This study is based on data collected from 300
forest stands located within 29 states. It was not
our intention in this paper to have an exhaustive
data set. Rather, we selected stands to represent
major community types within the deciduous
forest of eastern North America. Braun (1950)
was the source for 230 stands. All of the stands
selected from Braun (1950) contained at least 100
individual trees and represented actual stands
rather than regional averages. The sources for the
remaining stands included Quarterman & Keever
(1962) [29], Monk (1965) [26], Hartshorn (1972)
[4], Nesom & Treiber (1977) [3] and Sechrest &
Cooper (1970) [3], and Mills & Jones (1969) [1],
Monk (1961) [1], Nixon etal. (1980) [1], Skeen
(1974) [1] and Ware (1970) [1]. Most of the 70
stands not from Braun (1950) were from the
170
coastal plain, a region not well represented in
Braun's work. Stand distribution by states was as
follows: Kentucky 54, Michigan 43, Florida 29,
North Carolina 25, Pennsylvania 19, Tennessee
13, Indiana 11, Illinois 10, Virginia 9, Georgia 8,
New Hampshire 8, Ohio 8, West Virginia 8,
Louisiana 7, Maryland 7, South Carolina 7, New
York 6, Wisconsin 6, Alabama 4, Mississippi 3,
Arkansas 2, Iowa 2, Minnesota 2, Missouri 2,
Oklahoma 2, Texas 2, Connecticut 1, New Jersey
1 and Vermont 1.
87 woody taxa were recognized (Table 2). The
measure of species importance used in the analysis were relative density (Phillips 1959). These
Table 2. List o f species, the n u m b e r o f t i m e s e a c h r a n k e d 1,2 or 3 in 300 different stands. N u m b e r s in a n y c o l u m n m a y be g r e a t e r
t h a n 300 due to c o - d o m i n a n c e o f species.
Species
Fagus grandifolia
Acer saccharum
Quercus alba
Quercus rubra
Acer rubrum
Nyssa sylvatica
Fraxinus americana
Tsuga canadensis
Liriodendron tulipifera
Betula alleghaniensis
Castanea dentata
Liquidambar styraciflua
Tilia americana
Carya tomentosa
Ulmus americana
Carya spp.
Magnolia acuminata
Juglans nigra
Prunus serotina
Carya glabra
Quercus velutina
Quercus hemisphaerica
Quercus prinus
Betula lenta
Tilia heterophylla
Aesculus flava
Magnolia grandiflora
Pinus strobus
Quercus falcata
Carya ovata
Pinus taeda
Quercus nigra
llex opaca
Ulmus rubra
Total
stands
188
167
151
136
136
123
122
106
101
96
Dominance
level
1
2
3
65
35
35
3
2
37
37
29
14
8
34
10
2
2
20
19
20
18
23
16
19
23
3
7
14
20
21
81
17
6
5
76
76
68
65
62
60
59
52
52
47
43
11
8
2
1
-
13
8
11
3
4
8
13
10
11
9
2
3
3
1
8
5
3
1
10
1
2
7
3
41
41
5
3
3
2
40
40
40
38
38
37
36
34
33
26
4
6
8
4
6
4
-
6
1
5
3
2
1
4
2
1
1
7
8
7
2
4
3
4
4
1
2
Species
Carya cordiformis
Ostrya virginiana
Ulmus alata
Quercus michauxii
Picea spp.
Carpinus caroliniana
Comus florida
Quercus virginiana
Quercus stellata
Acer barbatum
Halesia carolina
Pinus glabra
Quercus muehlenbergii
Quercus coccinea
Abies spp.
Celtis occidentalis
Persea borbonia
Betula papyrifera
Pinus echinata
Quercus macrocarpa
Celtis laevigata
Quercus phellos
Pinus rigida
Quercus shumardii
Pinus resinosa
Quercus palustris
Quercus bicolor
Quercus ellpsoidalis
Quercus marilandica
Pinus virginiana
Acer saccharinum
Quercus imbricaria
Populus spp.
Total
stands
Dominance
level
1
2
-
2
11
-
1
2
3
19
19
15
1
1
4
3
1
15
15
1
1
3
-
4
14
14
13
-
2
3
2
2
3
12
-
1
1
9
8
7
6
6
5
5
4
3
3
3
3
3
2
2
-
1
1
2
1
1
-
-
2
2
1
1
1
1
1
1
24
24
21
21
20
20
19
1
1
1
3
2
2
13
1
3
3
5
2
1
2
1
1
1
2
1
1
1
-
1
-
O t h e r species a n d species g r o u p s recognized: Aesculus glabra, Betula nigra, Carya buckleyi (C. texana), Diospyros virginiana,
Fraxinus spp., Ilex spp., Juglans cinerea, Juniperus virginiana, (Jr. silicola), Magnolia spp., Morus rubra, Oxydendron arboreum,
Platanus occidentalis, Quercus spp., Q. lyrata, Sassafras albidum, Sabal palmetto, Taxodium distichum, Thuja occidentalis.
171
data were ordinated using detrended correspondence analysis (DCA) (Hill 1979a; Hill & Gauch
1980; Gauch 1982). The stand data were classified using a two-way indicator species analysis
(Hill et al. 1975; Hill 1979b). This analysis creates
simultaneously a classification of samples and a
classification of species according to their ecological preferences. The result is a tabular matrix
similar to a Braun-Blanquet table.
Recently, DCA has been criticized (Minchin
1987) but it is generally the most widely used and
cited ordination technique in ecological literature
and has distinct advantages for our application in
that stands and species are ordinated simultaneously and the computational requirements are
small. Also, our gradients are large relative to
most ordination applications which is an advantage of DCA over other methods (Gauch 1982).
As with any numerical analysis method, examination of assumptions and consideration of more
appropriate alternatives is important. DCA provides interpretable ordinations for our data set
which is likely the ultimate criteria for any multivariate method.
Results and Discussion
The forests used in this study are not evenly distributed within the eastern deciduous forest.
Since Braun (1950) was the source for 230 of the
300 stands, any bias which may have existed in
that source is retained. For example, about 60~o
of the individual stands given in Braun represent
either what she called mixed mesophytic forests or
the hemlock-white pine-northern hardwood
forests. The final selection of 230 stands from
Braun plus 70 stands from the piedmont and
coastal plain still resulted in an uneven distribution of stands among forest communities.
Composition of the eastern deciduousforest
Some species within the deciduous forest are
more wide-ranging than others, and some tend to
dominate stands in which they occur while others
are usually subdominant when present. 9 species
occurred in more than one-third of the stands
(Table2). Fagus grandifolia, Acer saccharum,
Quercus alba, Q. rubra, Acer rubrum, Nyssa sylvatica, Fraxinus americana, Tsuga canadensis and
Liriodendron tulipifera had the highest frequency.
Acer was the most frequently occurring genus,
followed by Quercus, Fagus, Fraxinus, Tilia,Nyssa
and Ulmus.
39 species ranked as the dominant at least once
(Table 2). Some species appeared as the leading
dominant more often than as subdominant (Fagus
grandifolia, Acer saccharum, Quereus alba, Tsuga
canadensis, Castanea dentata). Other species tend
to rank 2nd or 3rd in dominance rather than as 1st
(Betula alleghaniensis, Quercus rubra,Acer rubrum,
Carya tomentosa, Quercus velutina and Ulmus
americana). Other species, though frequent in
occurrence, were rarely dominant. This group
included Nyssa sylvatica, Fraxinus americana,
Carya spp., Magnolia acuminata, Juglans nigra,
Carya glabra, Prunus serotina, Aesculus flava and
Carya ovata.
Oaks (Quercus) are more important than
hickories (Carya) (Table 2). Despite references to
oak-hickory forests by many authors, 20 different
species of oak and only 5 species of hickory dominated at either the 1st, 2nd or 3rd level of dominance within the selected stands of this study.
Oaks ranked 1st, 2nd or 3rd 218 times as compared to 43 for the hickories. Therefore, most so
called oak-hickory forests have far more oak than
hickory and probably should be called mixed oak
or oak forests rather than oak-hickory forests.
For example, the William L. Hutcheson Memorial
Forest in New Jersey (Monk 1961) is a primeval
representative of what has been called an oakhickory forest. 47 percent of the trees and 86 ~ of
the basal area belong to the oaks while only 7 and
8 ~/o,respectively, belong to the hickories. Oosting
(1942) described oak-hickory forests found in the
piedmont of North Carolina that ranged from 65
to 80% oak and contained only 8 to 3 2 ~ hickory.
Ordination of the eastern deciduousforest
The DCA ordination of the stand provided interpretable patterns. Stand separation seems to cor-
172
c
ouercus
!
/
Acor
c
"
coastal Plain
8
s*~n_I~
s s
;~j
C
*
P
1%
p
C
c
I
Cc c~MUh
L
tlt
U
w~llt
w w w*v
w
39
d
o
t
I
I
J
200
6
I
I
I
400
v/
I
I
I
I
I
J
600
e
29
Fig. 1. DCA axis 1 and 2 locations for stands used in this
study. The symbolscorrespond to groupings recognizedusing
TWINSPAN (s and S = southern mixed hardwoods,
o=oak (Quercus), c=chestnut (Castanea), m and
M = mixed mesophytic, w = hardwoods, h = hemlock
(Tsuga), f = fir (Abies), p = white pine of the hemlock-white
pine northern hardwoods (Pinus). The solid line separates the
Acer Province from the Quercus Province. The dashed line
includes all of the coastal plain stands to the left; the dotted
line includes all of the Florida stands to the left.
2
M
e
a
2O
n
/%
n
n
u
1
respond to two complex-gradients (Whittaker
1975). Axis 1 corresponds to a latitudinal/elevational gradient (Fig. 1). Stands with spruce and fir
from higher elevations and latitudes are clustered
as a group on the right in Fig. 1, whereas stands
from the southeastern coastal plain are clustered
at the left. Moving from left to right the first set
of stands are from Florida; these grade into those
from the rest of the coastal plain which in turn
blend with those from the piedmont which finally
blend with stands from the foothills and
mountains (Fenneman 1938). All of the Florida
stands, with two exceptions, are positioned left of
the dotted line in Fig. 1. One Alabama and one
Mississippi stand are included in the Florida
group. Most of the other coastal plain stands are
positioned between the dotted and dashed lines in
Fig. 1. Several Virginia, North Carolina and
Georgia piedmont stands are included in this
grouping along with all of the bottomland
hardwood forests of the Mississippi drainage.
The several forests from the Missouri Interior
Highlands and Oklahoma Plateau also fall within
this coastal plain group. Several beech-dominated
forests from the coastal plain (Nesom & Treiber
10
T
III
m
P
Fig. 2. Relationship between latitude (A) and mean annual
temperature (°C) (B) with the axis 1 position from the DCA
ordination.
1977) were located within the non-coastal plain
stands.
The latitudinal range for the 205 stands was
29 ° to 40 ° (Fig. 2A). A significant positive
relationship existed between stand latitude (independent variable) and stand axis 1 score
(r 2 = 0.64, P < 0.001,205 df). There was no significant relationship between stand latitude and
axis 2 score (r 2 = 0.03). When mean annual temperature ( ° C) was used as the independent variable and axis 1 stand score as the dependent variable, a significant negative relationship was found
(r 2 = 0.74, P < 0.001,211 df). The mean annual
temperature ranges from 4 - 2 2 °C (Fig. 2B).
173
Axis 2 separation seems to correspond with
a complex-topographic-edaphic-gradient. The
species that dominate stands with low axis 2
scores and deciduous species usually associated
with mesic, fertile soils (Figs 1, 3, 4). Stands in
which evergreen species are dominant have high
first and/or second axis scores. Abies spp., Picea
spp., Tsuga canadensis and Pinus strobus dominate
stands with high axis 1 and axis 2 scores (Fig. 1).
Stands with intermediate axis 1 scores and high
axis 2 scores are dominated by Pinus rigida or by
a few deciduous species often associated with
nutrient-poor soils (Castanea dentata, Quercus
prinus or Q. coccinea).
Within the coastal plain cluster, stands dominated by evergreens (Magnolia grandiflora,
Quercus hemisphaerica, Q. nigra, Q. virginiana)
tend to have higher axis 2 scores than the
deciduous species (Acer barbatum, Celtis laevigata,
Liquidambar styraciflua). All of the pines that are
associated with secondary succession have high
axis 2 scores. Since the bottomland hardwoods
are located in the lower portion in Fig. 1, axis 2
may partially reflect a moisture gradient.
Species' positions in stand space are centroids.
Species with similar environmental requirements
occur close together while those with vastly different requirements are more distant from one
another (Gauch 1982). Further, the location of
each species represents that space where each
occurs with its highest percentage importance. As
one moves in any direction from that point the
Table 3. Abundance curves for 45 species as they are distributed along the x-axis in Fig. 1. Each column is equivalent to 25 x-axis
units. The 3-4 highest abundance values are underlined.
Species
(# Stands)
0uercus hemisphaerica
0uercus ~
Ouercus virginiarm
Persea borbonia
Ma~oolia ~ar~iiflora
Acerbarbatum
Pir~s taeda
Celtis laevigata
L i q u i d a ~ srvraciflua
Carvaglabra
~[y_@ tomentosa
Tilia floridana
Nvssa sylvatica
Ouercus stellata
Pinus glabra
Ouercus falcata
Ulmus alata
Carva cordiformis
Quercus _velutina
0uercus _marilandica
0uercus alba
Carva ovata
Quercus _¢occinea
Fraxinus americana
Ulmm ~rericana
Quercus prinus
F a ~ gremdifolia
Lirioden4ron tulipifera
Juglans
tqmus rubra
Maxilla acumir~ta
Quercus rubr_a
Castanea dentata
Tilia heterophvlla
Acer rubnJn
Aesculus flava
Betu]a lenta
Prur~s serotir~
Acer saccharun
Tilia americana
Tsu~a c~maensis
Betula alle~naniensis
Pinus strobus
Picea s-pp.
Abies spp.
1 2 3 4 5 6 7 8 9 i0 ii 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
3 1 15 1 13 1 4 7 6 14 16 7 13 19 6 13 23 27 21 26 16 ii 9 Ii 4 1 5 1 1 1 1 2
74
31
17
2
i0
1
i0
i
12
1
1
i
64 57 47 40 28 21 16 12 9 7 5 3 1
40 37 31 28 23 18 12 7 2
20 24 27 25 23 18 12 7
5 ii 13 13 4 i
20 28 50 52 51 49 45 40 34 28 23 20 16 II
52027323027242014
7 4 i
1 34 68 70 63 55 46 38 30 23 16 I0 4 i
I 5 23 28 18 7
13 24 37 47 48 49 47 44 42 37 33 25 17 8
6 81418232522171410
9 7 6 4
15 21 25 29 32 33 32 30 28 24 22 19 14 12
3 4 5 7 810121211
8 7 6 4 3
3 5 7 8 9 i0 I0 Ii i0 9 8 7 7 6
3 12 22 36 41 32 18 9 3 1
1 4 6 i0 12 15 17 19 18 16 15 13 II
3 7 12 17 23 28 33 30 20 16 12
i 3 6 8121516151311
9 7
i 3 5 6 7 8 12 15 17 20 17 14
8 30 42 37 33
1342621
5
3 8 12 21 26 37 45 64 79 80 81 76
1 4 5 6 7 8 9 ii 13 15 17 18
4 13 22 30 29 27
4 4 7 9 12 13 15 17 18 20 20 21
2 3 4 8 I0 ii 15 20 24 26 28 29 28
3 7 12 20 25 39 48 57 66
4 9 14 21 27 37 44 55 61 68 75 78
3 51014192532
146810
1 3
13
57
3 9 15 19 26
3 8 1 2 20
I 4
5 8 ii 12 14 16 18 21 23 24 23 22 22
2
2 3
3 4
7 3
i
i
3 4 5 6
i0 7 4 2
6 6
7 2
5 3 2
i
I
7 543
2
4 1
731
17 ii 7 2
57 44 36 28
ii 8 7 5
22 18 16 13
20 20 19 17
26 25 23 21
60 42 31 18
78 75 68 58
46 41 37 32
14141211
ii 14 14 13
9 i0 Ii 12
38 42 45 48
28 52 69 87
i0 14 18 23
22 23 24 25
5101624
1 61217
5 7 81013
12
34
5 6 8 1 0 14 27 42 51 63
1 3 5 810142030
4 16 20 31
2 6
i 4 8 12 16 22 30
I
21 13 8 4 1
3 1
8 3
15 ii i0 7 3
20 18 17 16 14 ii
II 4 1
494638302617
27 21 17 12 5 I
1063
1085
31
1096
5 3
48 33 21 15 7 3
83 57 26 14 8 3
26 27 16 6 2
26 28 26 23 21 20
302613 2
21 20 18 16 14 ii
15 17 18 15 12 7
72 80 87 89 79 65
38485447402915
48 58 65 73 83 88
i0 17 21 28 37 44
36 42 48 53 61 67
6 II 18 21 30 37
2 6 9
8
6
5 2
9 4
2 I
i
16 ii 8 4
9 4 2
5 2 i
53 42 27
4
78 55 37
47 45 38
71 73 75
45 52 60
12 15 18
i
16 II 3 i
24 12 4
28 17 7 2
68 73 77 80
21 23 22 22
174
R .04
•
3.
40
a
20
~
B
D
E
t
i
v
40
20
D
@
n
5
i
t
40
2O
Y
Fig. 3. Abundance of 13 species within the DCA ordination axis 1 and 2 matrix: (A) Quercus hemisphaerica, (B)Picea spp.,
(C) Magnolia grandiflora, (D) Tsuga canadensis, (E) Fagus grandifolia, (F) Quercus alba, (G) Q. prinus, (H) Castanea dentata,
(I) Liquidambar styraciflua, (J) Betula alleghaniensis, (K) Acer saccharum and (L) Pinus taeda and (M) P. strobus.
importance of the species declines (Table 3 and
Figs 3 and 4). Data for Table 3 were generated by
plotting the relative density of selected species
along their axis 1 score in species space. The
curves were smoothed by plotting species relative
density along axis 1, then drawing a curve that
encompassed all of the data points. The importance of each species was then read from the
smoothed curves. The species were arranged in
Table 3 so that those reached maximum importance in lower latitudes are listed first and those
with maximum importance at higher latitudes are
given last. The score on axis 1 where each species
reaches maximum density is underlined. Each
column represents 25 units on axis 1.
The turnover of species in Table 3 is suggestive
of an ecocline (Whittaker 1975). Certainly stand
position on axis 1 corresponds to a climatic
gradient and the change of species composition
represents a community gradient. No distinct
species groupings are obvious in Table 3. Rather,
a gradual change in species composition along
axis 1 occurs as one moves from top to bottom.
Twenty species range over at least one-half of axis
1 and these species give an element of floristic
continuity to the vegetation (Gleason 1926, 1939).
There is another group of 18 species not so
broadly distributed along the axis (one-third of
the total range). These species give a distinctive
element to each region. If one ignores the total
range of species distribution (Table 3) and concentrates on that portion where each species
reaches maximum density (underlined values);
overlapping distributions are still quite evident.
Using only the underscored values, 3 groupings
are evident. Axis 1 segments 1 through 11 are
separated from the rest except for the sharing of
high values of Quercus velutina, Q. marilandica,
Q. alba, and Carya cordiformis. Segment 1
through 11 roughly correspond to stands of this
coastal plain. A second break comes between axis
segments 24 and 25 where Tsuga canadensis domi-
175
2(]
R
e
1
a
t
i
v
i
E
40
2O
I
Y
J
"
L
Fig. 4. Abundance of 12 species within the DCA ordination axis 1 and 2 matrix: (A) Carya tomentosa, (B) Quercus virginiana,
(C) Tilia heterophylla, (D) Quercus falcata, (E) Q. marilandica, (F)Aesculusflava, (G)Liriodendron tulipifera, (H) Quercus rubra,
(I) Tilia americana, (J) Acer rubrum, (K) Quercus stellata and (L) .4hies spp.
nance bridges the gap. The third grouping separates segments 29 through 32 from the others.
Picea and Abies spp. are the dominants of the
latter group. This division of segments 1-11, 12-24
and 25-32 roughly corresponds to the 'southern
region', 'central hardwoods' and 'northern region'
of Mattoon (1936) or the southern maritime pine
belt, deciduous forest and northern pine belt of
Sargent (1884) (Table 1).
Dominant species relative densities are plotted
in stand space in Figs 3 and 4. Some species are
widely distributed with respect to DCA axes 1
and 2 (Fagus grandifolia (Fig. 3E), Quercus alba
(Fig. 3F), Acer saccharum (Fig. 3K), Acer rubrum
(Fig. 4J)) though none of these species are equally
important throughout the landscape or within the
ordination. Quercus alba, one of the most widely
distributed tree species reaches maximum importance in the left one-half of Fig. 3F. When it coexists with Castanea dentata (Fig. 3H) or Acer
saccharum (Fig. 3K), Quercus alba's abundance is
low.
Classification of the eastern deciduous forest
4 major groupings and several subgroupings were
produced by the TWINSPAN analysis (Table 4).
These are (1) hemlock-white pine-northern
hardwoods (Tsuga-Pinus), (2) beech-maple
forests (Fagus-Acer), (3) oak (Quercus) forests
and (4) southern mixed hardwoods. The hemlockwhite pine-northern hardwoods has two distinct
sub-groups. Subgroup 1 was dominated by Picea
spp., Abies spp. and Betula alleghaniensis (Table
4). Fagus grandifolia, Tsuga canadensis, Acer
rubrum, A. saccharum and Betula papyrifera were
next in importance. This grouping of stands represent those from high elevations in the Appalachian
Mountains or those of more northern latitudes
and they tie the eastern deciduous forest with the
subarctic-subalpine forest. Subgroup 2 contains
stands more typical of the hemlock-white pinenorthern hardwoods forest. Stands within
subgroup 2 may be dominated by Pinus strobus,
Tsuga canadensis or a variety of 'northern'
hardwoods such as Acer saccharum, Fagus grandifolia, Betula al'leghaniensis or Tilia americana
176
Table4. 8 TWINSPAN subgroups. Stands within
subgroupswerelumpedandmeanrelativedensitycalculated
for each species, t = trace.
........................
~ - ~ z - - - § ~ 7
No. Hwd.
Group
Species
~i~-~i~g~-
....
Maple
Group
i
2
3
4
(12) (51) (62) (32)
........
Abies spp.
Picea spp.
Betula alle2haniensls
Pinus strobu$
Tsuea canadensis
Betula l e ~
Acer saccha~lm
Aeer rubrum
Fa~us erandlfolia
Ulmus americana
Ulmus rubra
Tilla americana
Fraxinus americana
Ouercus r~bra
Ma~nolla acuminata
Aesculus flava
Tilia heteroDhvlla
Lirlodendron tuliDifera
Ju21ans ni~ra
Castanea dentata
Quercus Drlnus
Carva ovata
Ouercus coeeinea
Ouercus velutlna
Ouercus alba
Nvssa svlvatlca
Ouereus stellata
Carva tomentosa
Carva ~labra
Pinus taeda
Ouereus falcata
Ulmus alata
Liouldambar stvraelflua
Ouereus hemlsDhaeriea
Ma~nolla ~randlflora
Pinus ~labra
Ouercus ni~ra
Tilla florldana
Aeer barbatum
Ouereus vir~iniana
Celtls laevi2ata
Persea borbonla
~?~g~}-~[-§~[i~i ........
~
......
Group
~7-~i~-
Hardwoods
Group
5
6
7
(37) (41) (37)
8
(28)
~ ......................................
8
54
18
t
4
1
3
3
7
t
I
9
t
7
I
i
8
3
t
12
t
36
3
15
2
t
8
i
i
41
2
3
7
6
5
I
12
3
29
I
i
3
3
6
I
3
5
I0
i
2
t
i
t
t
5
2
2
I
7
3
I0
t
t
t
t
t
t
t
t
t
2
i
2
i
I
I
3
t
6
t
t
t
5
I
t
i
i
7
2
i
i
6
t
36
ii
t
I
t
7
2
3
i
4
t
i
1
3
5
t
5
1
i
2
2
2
6
35
I
3
2
i
6
2
t
t
i
t
I
21
I
i
t
2
t
t
I
t
t
2
12
3
2
5
2
6
5
i
I0
4
i
i
i
2
I
I
7
3
i
i
I
14
14
8
6
3
2
i
i
i
i
8
2
5
5
2
2
,,---,,---,,---,,---,,---,,---,,---,,-
(Table 4). The DCA separation of stands within
subgroup 2 placed Tsuga canadensis-dominated
stands (h) between those dominated by Pinus
strobus (p) and those dominated by the
hardwoods (w) (Fig. 1). Some of the stands within
the beech-maple (Fagus-Acer) forest and maplebasswood (Acer-Tilia) forest regions (Braun
1950) were included in subgroup 2. The beechmaple group was divided into two. Subgroup 3
represents mixed mesophytic forest stands with a
heavy dominance of Tsuga canadensis (M) and
subgroup 4 a grouping of more typical mixed
mesophytic forest stands (m). subgroup 4 has a
high level of dominance from Fagus grandifolia,
Acer saccharum and Liriodendron tulipifera.
Species of lesser importance were Quercus rubra,
Q. alba, Tsuga canadensis and Tilia heterophylla.
Stands of ravine bottoms and ravine slopes, such
as those of the Joyce Kilmer National Monument,
were mostly included in subgroup 3. The Tsugadominated (M) stands of the beech-maple
subgroup are closely grouped with the Tsugadominated (h) stands of the hemlock-white pinenorthern hardwoods. The stands which Braun
(1950) considered typical of the mixed mesophytic
forest fell within subgroup 4 along with about half
of those in her western mesophytic forest region.
Some of the stands within Braun's beech-maple
forest and maple-basswood forest regions also fell
within subgroups 3 and 4.
Many of the Castanea dentata dominated
stands were actually grouped in subgroup 4,
though most of them were positioned together in
that portion of the subgroup 4 that merged with
the Quercus dominated stands. Since the Castanea
dominated stands overlap completely with those
of subgroup 4 along DCA axis 1 but were separated along DCA axis 2 from subgroup 4 (Fig. 1),
they were removed from subgroup 4 and made a
subgroup under the oak group. Thus the Castanea
dominated stands (c) represent subgroup 5 and
the oak dominated stands (o) subgroup 6. Subdominants include Quercus prinus, Fagus grandi-
folia, Quercus alba, Q. rubra, Acer saccharum and
Liriodendron tulipifera. The oak subgroup 6 was
clearly dominated by Quercus alba, Q. velutina,
Q. rubra, Liriodendron tulipifera and Pinus taeda
represent subdominants. About half of Braun's
western mesophytic forest region stands were
included within subgroup 6.
The presence of Quercus hemisphaerica and
Liquidambar styraciflua separated the southern
mixed hardwoods group from the oak group. The
southern mixed hardwoods group was also
divided into 2 units. Subgroup 7 stands (S) have
a high level of importance of Fagus grandifolia,
Quercus alba and Liquidambar styraciflua.
Magnolia grandiflora, Quercusfalcata, Pinus taeda
and Carya tomentosa were important associates.
Most of the stands used from Quarterman &
Keever (1962) fell within subgroup 7 along with
other coastal plain stands with Fagus grandifolia
and Quercus alba. Subgroup 8 stands (s) have
177
little or no Fagus grandifolia or Quercus alba.
Leading dominants were Quercus hemisphaerica
and Liquidambar styraciflua with moderate levels
of Quercus nigra, Magnolia grandiflora, Carya
tomentosa, Quercus virginiana and Acer barbatum.
All of the southern mixed hardwoods forest
stands from Monk (1965) were included in
subgroup 8 along with other stands from the lower
coastal plain.
Species richness (total number) was lowest in
the hemlock-white pine-northern hardwoods
subgroups 1 and 2 and particularly so in those
stands dominated by Picea and Abies spp. Diversity was also rather low in the Castanea subgroup
5 and the beech-maple subgroup 3, dominated by
Tsuga canadensis. Species richness in all of the
other subgroups was much higher (Table 4).
Classification within Quercus and Castanea dominated forests
We did not expect the Castanea subgroup to be
similar to the beech-maple subgroup. We were
also surprised to find the oak subgroup containing
only 41 of the 130 stands in which a Quercus
Table 5. 6 TWlNSPAN subgroups from 154 stands dominated by Quercus and Castanea species. The relative density
of the leading dominants for each subgroup are underlined.
t = trace; ( ) number of stands in sub-groups.
..................................................
Species
Sub-groups
i
2
3
4
5
6
(12) (41) (38) (23) (14) (26)
~oii;-~ran~i~iora
.......
Ouercus hemisD~aerica
Quercu~ vlr~inlana
Ouercus ni~ra
Pin~ ~d a
Ouercus falcata
Liauidambar styraciflua
Ouercus
stellata
CarTa ~labra
Carya tomentosa
Nvssa svlvatica
Ouercus velutina
Ouercus cocclnea
Ouercus Drlnus
Ouercus alba
FaKus ~randlfolia
Lirlodendron tuliDifera
Fraxinus amerl~ana
Acer rubrum
Castanea denta~a
Carya ovata
Quevcu$ ~ubra
T i l ~ americana
Aeer saccharum
Ouercus macrocarDa
~---g ..................
12
7
7
I
1
ii
4
8
2
2
1
2
2
i
Ii
2
1
12
4
9
1
8
7
2
i
9
9
2
I
1
6
3
6
6
1
3
4
7
5
16
13
4
1
3
i
2
1
t
2
3
4
6
3
21
6
9
4
2
2
4
6
i
5
t
1
1
3
1
1
5
8
12
8
2
5
i~
1
8
i
7
9
2
2
9
I
2
i
19
16
18
2
species occurred as one of the 3 leading dominants. To better understand these relationships,
154 stands in which Quercus or Castanea were
leading dominants were selected from the 300
initial stands and new ordinations (DCA) and
two-way classification (TWlNSPAN) were run.
On the basis of these analyses, 6 subgroups
(Table 5) were identified. Subgroup 6 was dominated by Quercus rubra, Tilia americana and Acer
saccharum, Quercus alba and Fraxinus americana
were subdominants. Subgroup 5 had Castanea
dentata and Fagus grandifolia as the leading dominants with Quercus rubra, Q. alba and Liriodendron tulipifera as subdominants. Quercus alba
was the clear dominant in subgroup 4 with a
relative importance twice that of any other
species. This was the most species-rich subgroup.
Subgroup 3 had Q. alba and Fagus grandifolia as
the leading dominants. The oaks of drier habitats
(Quercus coccinea, Q. stellata, Q. velutina) and
some successional species (Pinus taeda) were
important subdominants. Southern mixed
hardwoods comprised subgroup 1 and subgroup
2 was transitional between it and subgroup 3. The
only real difference between the two ordination
results was the elimination of the hemlock-white
pine-northern hardwoods stands from the 154
oak data set.
When stands ordinated with DCA are identified to TWlNSPAN groupings, they form compact clusters (Fig. 1). The hemlock-white pinenorthern hardwoods cluster has two extending
arms; one that represents Abies-Picea (f) stands of
higher elevations and one with Pinus strobus (p)
probably reflecting past disturbance and succession. The Tsuga-dominated stands (h) separate the Abies-Picea and Pinus stands from the
hardwood-dominated ones. The hardwood
subgroup (w) merge with the beech-maple
subgroup (m). The Castanea subgroup (c) is distinctly separated from the beech-maple subgroup.
The beech-maple group gives way rather abruptly
into the oak group (o) which blends more gradually into the southern mixed hardwoods (s). There
is minimal overlap between the oak group and the
Castanea subgroup.
The gradient of species illustrated in Figs 3
178
and 4 and Tables 3-5 are suggestive of an ecocline
(Whittaker 1975) defined as a combination of a
complex environmental gradient with a community gradient (coenocline). Stand positions
follow a complex gradient that more or less corresponds to a combination of latitude/elevation
and temperature (Fig. 2).
Classification summary
A problem that has plagued ecologists is whether
the deciduous forests of eastern North America
can be divided or classified into natural entities.
TWINSPAN may provide an objective means to
answer this question. The 300 initial stands are
divided into 2 groups based on the presence of
Liquidambar styraciflua-Quercus hemisphaerica
and the absence ofAcer saccharum. At the second
division level the stands with Acer saccharum are
divided into those with Betula alleghaniensis and
the absence of Quercus alba-Liriodendron tulipifera
and those without Betula alleghaniensis and the
presence of Quercus alba-Liriodendron tulipifera.
The Liquidambar styraciflua-Quercus hemisphaerica stands are divided at the second level into
those with Fagus grandifolia-Quercus alba and into
Liquidambar styraciflua-Quercus hemisphaerica
stands without Fagus grandifolia-Quercus alba. At
this level of classification one begins to fmd many
of the same species present in several stand
groupings.
Species are not equally valuable as community
indicators (Daubenmire 1966). With this in mind
28 taxa were omitted as indicator species and the
300 stands reclassified by TWINSPAN. The
omitted taxa were Abies spp., Acer spp.,
A. rubrum, Carya spp., C. buckleyi, Celtis occi-
dentalis, Diospyros virginiana, Fagus grandifolia,
Fraxinus spp., F. americana, Ilex spp., Magnolia
spp., Morus rubra, Picea spp., Pinus banksiana,
P. echinata, P. resinosa, P. rigida, P. virginiana,
P. taeda, Populus spp., Quercus spp., Q. ellipsoidalis, Q. imbricaria, Sassafras albidum, Thuja oceidentalis and Ulmus americana. The reclassification
was basically the same. The 300 stands were separated into those with Liquidambar styraciflua
-Quercus hemisphaerica and the absence of Acer
saccharum-Quercus rubra. In this classification
Quercus rubra now becomes an indicator species.
Most of the important tree species in the
eastern deciduous forest have wide geographical
ranges. Efforts to discern classifiable units generally fail to provide distinct classes. Oosting (1956)
described the beech-maple forest as a northern
extension of the mixed mesophytic forest and
stated that Fagus grandifolia was replaced by Tilia
americana to form the maple-basswood forest.
Even Braun (1950) described forest stands with
species composition differing from the forest
region in which they were located. Tables 4 and
5 illustrate why this happens. Nearly half of the
tree species listed occurred in more than half of
the subgroups. Those listed include important
dominants as Acer saccharum, A. rubrum, Betula
alleghaniensis, B. lenta, Castanea dentata, Carya
glabra, C. tomentosa, Fagus grandifolia, Fraxinus
americana, Liriodendron tulipifera, Nyssa sylvatica,
Quercus alba, Q. coccinea, Q. prinus, Q. rubra,
Q. velutina, Tilia americana and Tsuga canadensis.
Perhaps complexity of species composition within
the eastern deciduous forest is more important
than the classifiability of stands.
Daubenmire (1978) placed the eastern deciduous forest in the Temperate Mesophytic Forest
Region and divided it into two provinces: (1)Acer
saccharum Province and (2)Quercus falcata
Province. The Acer saccharum Province more or
less coincides with the distribution of Acer
saccharum and it includes three sections. The
Quercusfalcata Province roughly corresponds to
the distribution of that species; New Jersey, into
peninsular Florida, west of the Appalachian
Mountains from Kentucky and Missouri to the
Gulf of Mexico, including eastern Texas.
Daubenmire (1978) recognized three sections.
The division of the eastern deciduous forest into
the A cer saccharum and Quercus falcata Provinces
is simply not adequate. Acer saccharum occurs in
many stands within the Quercusfalcata Province.
Perhaps more importantly Q.falcata is not an
important component in most forests within the
province. Often it is not present at all. The
eastern deciduous forest could be designated the
179
Acer-Quercus Forest Region. Such a designation
might also be an oversimplification for Quercus
species are important through much of the region
that would be designated Acer and Acer species
range through much of the area that would be
called Quercus.
If the deciduous forest has a distinct identifiable
subunit, it would be the southern mixed
hardwoods, which covers the southeast coastal
plain. The TWINSPAN classification of the 300
stands into those with Acer saccharum and
Quercus rubra and those with Liquidambar
styraciflua and Quercus hemisphaeriea mostly
separates the southern mixed hardwoods from the
rest of the deciduous forest. The distribution of
the southern mixed hardwoods closely tracks the
distribution of Q. hemisphaerica. The autecology
of Liquidambar styraciflua may be confusing as it
is often associated with bottomland forests, in the
coastal plain. However, in upland positions in the
coastal plain L. styraciflua also is a secondary
successional species (Monk 1968). Most stands
from the coastal plain used in this study were
probably late successional. A number of other
species tie the southern mixed hardwoods with
the rest of the deciduous forest (Carya glabra,
C. ovata, C. tomentosa, Quercus alba, Q. falcata,
and Fagus grandifolia to name a few).
Acer saccharum is important as an indicator
species; the initial TWINSPAN division of the
300 stands into two groups employed Acer
saccharum. Further subdivisions of stands within
theAcersaccharum group is more difficult. Several
combinations of species occur together frequently
enough to constitute subgroups. These stands
most often occur close to one another in the
landscape (Fig. 1) and again probably reflect similar environmental requirements of species rather
than interspecific ties. For example, TWINSPAN
separated stands into several recognizable
subgroups based on the dominance of Picea spp.,
Pinus strobus, Tsuga canadensis, Tilia americana,
Fagus grandifolia, Castanea dentata and Quercus
alba. Even though there are many stands dominated by a few species, they are not distinct
enough to be clearly demarcated. The more ordination axes used to separate the groups, the more
distinguishable they become (Fig. 1). There is so
much overlap in species distributions that only
regional trends are of much use.
The Society of American Foresters divided the
forests of eastern North America into three forest
regions (1)northern, (2)central and (3)southern
(Eyre 1980). Using basal area of the leading dominants, 62 forest cover types were recognized. The
cover types are based on what is present today. In
the 300 stands employed here 12 species are
among the leading 3 dominants in at least 20 of
the stands (Table 2). The 12 are among the most
widespread species in this study (Acer rubrum,
A. saccharum, Betula alleghaniensis, Carya tomentosa, Castanea dentata, Fagus grandifolia, Liquidambar styraciflua, Liriodendron tulipifera, Quercus
alba, Q. rubra, Tiliaamericana and Tsuga canadensis) and they all are important species in the SAF
cover types. Each cover type occupies a rather
large area that may not be in continuous stands.
They may occur sporadically and merge with
others. 'Throughout much of the eastern
hardwood forest, particularly on good sites, the
seemingly chaotic mingling of species often makes
type determination difficult' (Eyre 1980).
The deciduous forest could be called the
Quercus-Acer Forest Region. Acer saccharum
probably has more indicator value than Fagus
grandifolia. The Acer portion of the region has
two sections (1) hemlock-white pine-northern
hardwoods and (2) beech-maple. The Quercus
portion of the region has three sections:
(1) Castanea dentata, (2) Quercus alba and
(3) Q. hemisphaerica sections; however, the
reduction of Castanea dentata by the chestnut
blight has effectively eliminated that entity. The
Castanea dentata section today is mostly dominated by Quercus prinus which blends strongly
with Q. alba and should probably be considered
as part of the Quercus alba section. This would
lump all of the non-coastal plain Quercus
stands. The Q. hemisphaerica section is distributed over the coastal plain from southeastern
Virginia to eastern Texas, though it may only
occur on a small proportion of the area. Most
sites actually have early successional stands
and/or fire-maintained communities that have
180
never supported the Q. hemisphaerica section
(Christensen 1988; Veno 1976; Monk 1968).
One could effectively argue that the division of
the eastern deciduous forest into four subunits
within the Acer-Quercus Forest Region is too
simple for such a complex forest system. Others
might ask why not use Fagus-Quercusrather than
Acer-Quercus? At the geographical scale of the
eastern deciduous forest, the system is not clearly
separable into a few clearly defined units. Perhaps
a realistic scheme could be derived by combining
portions of the various classifications offered in
Table 1. Most current authors consider the hemlock-white pine-northern hardwoods (Northern
Pine Belt of Sargent 1884, Lake Forest of
Clements 1928, Northern Region of Mattoon
1936) a part of the eastern deciduous forest rather
than as a separate system. The southern mixed
hardwoods (Southern Maritime Pine Belt of
Sargent 1884, Southern Region of Mattoon 1936)
is the most distinctive unit within the eastern
deciduous forest. That leaves the deciduous forest
of Transeau (1905) or the Central Hardwoods of
Mattoon (1936) which Clements (1928) divided
into maple-beech, oak chestnut and oak-hickory.
Braun (1950) would suggest the Central
Hardwoods be divided into maple-basswood,
beech-maple, mixed mesophytic, western mesophytic, oak-chestnut, oak-hickory and oak-pine.
Oosting (1956) suggested the maple-basswood
and beech-maple represented less diverse
northern extensions of the mixed mesophytic
forest. The loss of chestnut from the oak-chestnut
forest has prompted some to lump the oakchestnut with the oak-hickory forests (Barbour
etal. 1987). Following this line of thought the
eastern deciduous forest would be divided as:
1. Northern Forest Region
a. Hemlock-White Pine-Northern
Hardwoods
2. Central Forest Region
a. Beech-Maple
b. Oak-Hickory
3. Southern Forest Region
a. Southern Mixed Hardwoods
The Acer Province would include 1 and 2a while
the Quercus Province would include 2b and 3.
Acknowledgements
The authors wish to thank Drs N.L. Christensen,
E.E. Clebsch; W.H. Martin and S. Ware for their
reviews of early versions of this manuscript. We
would like to thank J. Aber for initially suggesting
this analysis.
References
Barbour, M.G., Burk, J.H. & Pitts, W.D. 1987. Terrestrial
plant ecology. 2nd ed. Benjamin/Cummings Publishing Co.
Braun, E.L. 1950. Deciduous forests of eastern North
America. Blakiston Co., Philadelphia.
Christensen, N.L. 1988. Vegetation of the southeastern
coastal plain. In: Barbour, M.G. & Billings, W.D. (eds),
North American terrestrial vegetation, pp. 317-363.
Cambridge University Press.
Clements, F.E. 1916. Plant succession. Carnegy Inst. Wash.
Publ. 242, 1-512.
Clements, F.E. 1928. Plant succession and indicators. H.W.
Wilson Co. New York.
Curtis, J.T. 1959. The vegetation of Wisconsin; An ordination
of plant communities. University of Wisconsin Press,
Madison.
Daubenmire, R. 1966. Vegetation: identification of typical
communities. Science 151: 291-298.
Daubenmire, R. 1978. Plant geography with special reference
to North America. Acad. Press, New York.
Eyre, F.H. (ed.) 1980. Forest cover types of the United States
and Canada. Soc. Am. For. Washington, D.C.
Fermeman, N.M. 1938. Physiography of eastern United
States. McGraw Hill Book Co. New York.
Gauch Jr, H.G. 1982. Multivariate analysis in community
ecology. Cambridge Univ. Press, Cambridge.
Gleason, H.A. 1926. The individualistic concept of the plant
association. Bull. Torr. Bot. Club 53: 7-26.
Gleason, H.A. 1939. The individualistic concept of the plant
association. Am. Midl. Nat. 21: 92-110.
Gleason, H.A. & Cronquist, A. 1964. The natural geography
of plants. Columbia Univ. Press, New York.
Greller, A.M. 1988. Deciduous forest. In: Barbour, M.G. &
Billings, W.D. (eds.), North American terrestrial vegetation, pp. 287-316. Cambridge University Press.
Hartshorn, G.S. 1972. Vegetation and soil relationships in
southern Beaufort County, North Carolina. J. Elisha
Mitchell Sci. Soc. 88: 226-238.
Hill, M.O. 1979a. DECORANA-A FORTRAN program for
181
detrended correspondence analysis and reciprocal averaging. Ithaca, N.Y., 52 pp.
Hill, M.O. 1979b. TWINSPAN-A FORTRAN program for
arranging multivariate data in an ordered two-way table by
classification of the individuals and attributes. Ithaca,
N.Y., 89 pp.
Hill, M.O., Bunce, R.G.H. & Shaw, M.W. 1975. Indicator
species and analysis, a divisive polythetic method of classification and its application to a survey of native
pinewoods in Scotland. J. Ecol. 63: 597-613.
Hill, M.O. & Gauch Jr, H.G. 1980. Detrended correspondence analysis, an improved ordination technique.
Vegetatio 42: 47-58.
Ktichler, A.W. 1964. Potential natural vegetation of the conterminous United States. Am. Geogr. Soc. Spec. Publ. 36:
1-155.
Mattoon Jr, W.R. 1936. Forest trees and forest regions of the
United States. U.S. Dept. Agr. Misc. Publ. 217, 1-55.
Mills, R.H. & Jones Jr, S.B. 1969. The composition ofa mesic
southern mixed hardwood forest in south Mississippi.
Castanea 34: 62-66.
Minchin, P.R. 1987. An evaluation of the relative robustness
of techniques for ecological ordination. Vegetatio 69:
89-107.
Monk, C.D. 1961. The vegetation of The William L.
Hutcheson Memorial Forest, New Jersey. Bull. Torr. Bot.
Club 88: 156-166.
Monk, C.D. 1965. Southern mixed hardwood forest of north
central Florida. Ecol. Monogr. 35: 335-354.
Monk, C.D. 1966. An ecological significance of evergreenness. Ecology 47: 504-505.
Monk, C.D. 1968. Successional and environmental relationships of the forest vegetation on north central Florida. Am.
Midl. Nat. 79: 441-457.
Monk, C.D., McGinty, D.T. & Day Jr, F.P. 1985. The
ecological importance of Kalmia latifolia and Rhododendron
maximum in the deciduous forest of the southern
Appalachians. Bull. Torr. Club. 112: 187-193.
Nesom, G.L. & Treiber, M. 1977. Beech-mixed hardwoods
communities: A top-edaphic climax on the North Carolina
coastal plain. Castanea 42:119-140.
Nixon, E.S., Marietta, K.L., Littlejohn, R.O. & Weyland,
H.B. 1980. Woody vegetation of an American Beech (Fagus
grandifolia) community in eastern Texas. Castanea 45:
171-180.
Oosting, H.J. 1942. An ecological analysis of the plant com-
munities of Piedmont, North Carolina. Am. Midl. Nat. 28:
1-126.
Oosting, H.J. 1956. The study of plant communities. 2nd ed.
W.H. Freeman & Co., San Francisco.
Parker, A.J. 1985. Compositional gradients in mesophytic
forests of eastern North America. Phys. Geog. 6:
247-259.
Parker, A.J. 1987. Structural and functional features ofmesophytic forests in the east-central United States. Ass. Am.
Geogr. 77: 423-435.
Polunin, N. 1960. Introduction of plant geography.
McGraw-Hill Co., New York.
Quarterman, E. & Keever, C. 1962. Southern mixed
hardwood forest: climax in the southeastern Coastal Plain.
Ecol. Monogr. 32: 167-185.
Robiehand, B. & Buell, M.F. 1973. Vegetation of New Jersey;
A study of landscape diversity. Rutgers University Press,
New Brunswick, NJ.
Sargent, C.S. 1884. Report on the forests of North America,
exclusive of Mexico. 10th Census. Vol. 9. Washington.
Sechrest, C.G. & Cooper, A.W. 1970. An analysis of the
vegetation and soils of upland hardwood stands in the
piedmont and coastal plain of Moore County, North
Carolina. Castanea. 35: 26-57.
Shreve, F. 1917. A map of the vegetation of the United
States. Geogr. Rev. 3: 119-125.
Skeen, J.N. 1974. Composition and biomass of tree species
and maturity estimates of a suburban forest in Georgia.
Bull. Torr. Bot. Club. 101: 160-165.
Transeau, E.N. 1905. Forest centers of eastern America. Am.
Nat. 39: 875-889.
Veno, P.A. 1976. Successional relationships of five Florida
plant communities. Ecology 57: 498-508.
Walter, H. 1985. Vegetation of the earth and ecological systems of the geo-biosphere. 3rd ed. Springer-Verlag, New
York.
Ware, S.A. 1970. Southern mixed hardwood forest in the
Virginia coastal plain. Ecology 51: 921-924.
Ware, S.A. 1982. Polar ordination of Braun's mixed mesophytic forest. Castanea 47: 403-407.
Weaver, J.E. & Clements, F.E. 1938. Plant ecology. 2nd ed.
McGraw Hill Book Co., New York.
Whittaker, R.H. 1956. Vegetation of the Great Smoky
Mountains. Ecol. Monogr. 26: 1-80.
Whittaker, R.H. 1975. Communities and ecosystems. 2nd ed.
MacMillan Publ. Co., New York.