VEGETATION ANALYSIS, ENVIRONMENTAL RELATIONSHIPS,

AND POTENTIAL SUCCESSIONAL TRENDS IN THE

MISSOURI FOREST ECOSYSTEM PROJECT

Stephen G. Pallardy'

Abstract: The vegetation data set of the Missouri Forest Ecosystem Roject (MOFEP, initiated by the Missouri
Department of Conservation) in the Ozark Mountains of southeastern Missouri was ordinated by Detrended
Correspondence Analysis (DCA) to identify vegetation gradients and potential environmental influences. The first
DCA axis represented a strong gradient in black oak importance and segregation of this species from white ash, black
walnut, eastern redcedar, sugar maple and bitternut hickory. Some Spearman rank correlation coefficients between
Axis 1 and measured environmental variables were significant but not high, suggesting that unmeasured factors (e.g.,
disturbance history, soil pH) were associated with species distribution along the first axis. The second axis clearly
separated mesic species (bitternut hickory, black walnut, white ash and sugar maple) from xeric species (shortleaf
pine, eastern redcedar, post oak and blackjack oak). Aspect and aspect x slope variables correlated strongly with Axis
2 scores and species importance, suggesting that this axis represented a habitat moisture gradient. Potential canopy
replacement in MOFEP forests was examined by analysis of overstory-understory relationships and density us. size
class distributions. White oak likely will become more important in MOFEP forests with time, as this species is
relatively abundant in the understory, while relative abundance of red oak group species and shortleaf pine is
substantially reduced compared to that in the overstory.

INTRODUCTION

The Missouri Forest Ecosystem Project (MOFEP) in the Ozark Mountains of southeastern Missouri represents one of
the most extensive and comprehensive long-term ecological studies undertaken in the eastern United States
(Brookshire and Hauser 1993). There have been few quantitative studies of the vegetation in the region and none of
this extent ( Nigh and others 1985, Read 1952, Ware and others 1992, Zimmerman and Wagner 1979). The baseline
vegetation data set collected for this project offered an opportunity to provide an initial, coherent view of the nature
and condition of this ecosystem and its plant population dynamics. Specifically, this study was undertaken to provide
an ecological analysis of MOFEP forests with regard to tree species composition and dominance, vegetation-
environment relationships, and potential successional trends.

METHODS

Data Analysis

The nine MOFEP study sites are located in Carter, Ripley and Shannon counties in the southeastern Missouri Ozarks
and range in size from 260 and 527 ha. Study areas, vegetation plot design and data collection were described in
detail by Brookshire and Hauser (1993). For the purposes of the present study, data consisted of 644 0.2 ha circular
"tree" plots on which species and diameter of all individuals 21 1.4 cm (4.5 in) dbh were recorded. Additionally,
within the boundary of each tree plot, species and diameter were recorded on four nested 0.02 ha plots for woody

'School of Natural Resources, 1-31 Agriculture, University of Missouri, Columbia, Missouri 65211.

551 10th Central Hardwood Forest Conference

plants ("saplings") that were between 3.8 and 11.3 cm (1.5-4.5 in) dbh. Within each 0.02 ha plot, "seedling" data
were recorded by species on a concentric 0.004 ha plot for plants > 1 m tall and < 3.8 cm (1.5 in) dbh.

I n d i i t gradient analysis of overstory species was conducted using tree plot data and Detrended Correspondence
Analysis @CA) (Hill and Gauch 1980). Tree Importance Value (IV= relative density + relative basal area) by
species for each plot was used as an indicator of abundance. As employed in this study, output of the DCA
procedure provides gradients in overstory species composition in two or more dimensions. Composition of major
forest types on MOFEP study sites was assessed by averaging IV data across plots of the same Ecological Land Type
(ELT) (Miller 1981). TWINSPAN analysis (Hill 1979) of the MOFEP data set also was conducted, but provided no
detectable advantage in forest classification over that offered by the ELT system (data not shown).

The relationships among DCA axis scores, species importance value and environmental factors were explored by
Spearman rank correlation analysis using the CORR procedure of SAS (SAS Institute 1988). Correlation analysis
allowed identification of: (1) species that contribute to vegetation gradients extracted in the ordination process and (2)
environmental factors that may influence species distribution and dominance across ordination gradients. Spearman
rank correlation was used instead of Pearson linear correlation to free the analysis of assumptions of linearity in
relationships between species importance and environmental factors and of normality in distribution of vegetation
across the ordination gradients (Dollar and others, 1992). Environmental data were obtained from data sets supplied
by the Missouri Department of Conservation that contained information for each 0.2 ha plot indicating slope %, aspect
(" azimuth) and slope position (I= floodplain, 5=ridgetop). For correlation analysis, aspect was transformed
lrigonometrically using a function that provided a maximum value (+I) for NE aspects (45") and minimum value (-1)
for SW aspects (225") (Beers and others 1966):

Transformed aspect = COS (Aspect - 453

This transformation provides a representation of the "mesicness" continuum of topographic orientation, with NE
aspects providing the most "mesic," SW aspects the most "xeric" environments, and other aspects being assigned
intermediate values along this microenvironmental continuum. Another variable was derived from transformed aspect
by multiplying it by slope % based on the tendency of microclimatic influences to be accentuated by the slope of the
land (Frank and Lee 1966).

Potential successional trends in these forests were investigated through overstory-understory and size class analysis on
groups of plots segregated by ELT classification. Data from the overstory plots mentioned previously and those of
the four 0.02 ha sapling plots were merged, and differences between understory k19.05 cm dbh) and overstory
(>19.05 cm dbh) relative densities of important species of each ELT were assessed by t-tests after arcsin square root
transformation ( Dollar and others 1992, Sokal and Rohlf 1981).

Plot density (#/ha) of seedlings (from 0.004 ha plots), saplings and trees (as mentioned above) was assembled and
segregated by ELT, and data for each species were plotted by diameter class [< 1.5 in (3.8 cm), 1.5-4.5 in(3.8-11.4
cm), 4.5-7.5 in (11.4 cm-19.1 cm),7.5-10.5 in (19.1-26.7 cm), 10.5-13.5 in (26.7-34.3 em), 13.5-16.5 in (34.3-41.9
cm), 16.5-19.5 in (41.9-49.5 cm), 19.5-22.5 in (49.5-57.2 em), 22.5-25.5 in (57.2-64.8 cm), > 25.5 in (64.8 cm) dbhl
for further analysis of population structure and potential replacement of canopy trees (Harcombe and Marks 1978,
Harcombe 1987, Johnson 1992, McCarthy and Wistendahl1988). Density vs. size class relationships for different
species were compared visually and statistically (using x2-basedcomparison of frequency distributions across size
classes) (Sokal and Rohlf, 1981).

letb Ccntrnl Hardwood Forest Confenace 552

 RESULTS AND DISCUSSION

Ordination and Classification

Seventy percent of the variation in vegetation composition was displayed in the first two (of four total) DCA
ordination axes. The sample ordination graph, using the first two axes, displayed a dense cloud of plots toward the
lower left-hand corner, with a far more scattered and sparse distribution of plots to the right and upper portions of the
graph Fig. 1). The DCA species ordination graph (Pig. 2) and correlation analysis of DCA plot scores with

Species Ordination
so0
Sample Ordination Cc
400

wo -
400

roo
-
So
•
c g m ~ s c,
J"
.
u&'&
Fo. Omuo
•

4::.
Ctom
Ol roo - 00. l
.**
(Y

.-
U)
200-

e
.-
V1
$ loo-
Ov
Cte* Oc
uo
2 loo. Po
l
0.
- 0.

.
0 0
a :
e
*
Omor
0 - t -100 - Jv

-200
-100 -100 0 100 200 300 400 M O
-so 0 W 100 1W 200 2% 300 S I O
Axis 1 Axis 1
Figure 2. Scatter plot of species across two-dimensional
Figure 1. Distribution of plots across two-dimensional space fmmDCA ordination of
ordination space derived from DCA ordination of tree
importance values of 644,0.2 ha plots.
values. Ley: Pe= shortled pine; .a
scarlet oak; Qv= black oak; Qs=post oak; Qa= white oak;
Qmue= chinkapin oak; Qmar= biackjack oak; Ns= black
species IV vable 1) indicated that the fit DCA axis ~ mCF; pignut hickory; C e bitternut hickory; Ctex.:
represented a strong gradient in black oak (Quercus black hickory; Ctom=mockernut hickory; Sa= sassafras; Cf=
velutha Lam.) importance and segregation of this flowering dogwood; Jv= eastern redcedar; Ur= slippery elm;
species from species such as white ash (Fraxinus Ua= winged elm, Fa= white ash; As= sugar maple; Jn=
americana L.), black walnut (Juglans nigra L.), black walnut.
eastern redcedar (Juniperus virginiana L.), sugar
maple (Acer saccharurn Marsh.) and bitternut hickory
(Carya cordifomis (Wangenh.) K. Koch). Black oak was more abundant in plots with low Axis 1 scores while the
latter group of species was important in plots with high Axis 1 scores. In a similar fashion, the second DCA axis
clearly separated mesic species such as bitternut hickory, black walnut, white ash, and sugar maple (high Axis 2
scores) from xeric species such as shortleaf pine (Pinus echimta Mill.), eastern redcedar, post oak (Quercus ~tellata
Wangenh.) and blackjack oak (Quercus man'landica Muenchh.) (low Axis 2 scores).

When Axis 1 scores were correlated with environmental factors, some relationships were apparent, with the strongest
correlations shown between axis scores and slope position (Table 2A). In practical terms, this correlation indicated
that black oak was more important on ridgetop positions. However, in general, environmental correlations with Axis 1

553 10th Central Hardwood Forest Conference

Table 1. Spearman correlation coefficients between DCA axis scores and species importance values. ** pd.01;
* pfl.05

Ordination Axis

Axis 1 Axis 2

ACERSAC*
CARYCOR
CARYGLA
CARYTEX
CARYTOM
CORNFLO
FRAXAME
JUGLNIG
NNMR
wsssn
PINUECH
QUERALB
QUERCOC
QUERMAR
Q U E W
QUERSTE
QUERVEL
SASSALB
ULMUALA
ULMURUB

*Species codes for Tables 1,2 and 3. ACERSAC = sugar maple; CARYCOR = bitternut hickory; CARYGLA =
pignut hickov, CARYTElX = black hickory; CARYTOM = mockernut hickory; CORNFLO = flowering dogwood,
FRAXAME = white ash; JUGLNIG = black walnut; JUNIVIR = eastern redcedar, NYSSSYL = black gum;
PINUECH = shortleaf pine; QUERALB = white oak; QUERCOC =scarlet oak; QUERMAR = blackjack oak;
QUERMUE = chinkapin oak; QUERSTE = post oak; QUERVEL = black oak; SASSALB = sassafras; ULMUALA =
winged elm, ULMURUB = slippery elm.

were not strong (i.e., less than 0.3), suggesting that some unmeasured factor likely was important in determining
species composition across this axis. Plausible candidates are disturbance history and soil factors (especially pH, Nigh
and others 1985, Ware and others 1992). In contrast to these relatively weak Axis lenvironment correlations, the
transformed aspect and aspect x slope variables correlated more strongly (r = 0.43 to 0.46) with Axis 2 DCA sample
scores. This result is consistent with the distribution of species shown in the species ordination graph (Fig. 2). Hence
Axis 2 of the DCA ordination appeared to represent a habitat moisture gradient.

MOFEP Forest Composition

Forests of the two most common ELT classifications (17 and 18) (Table 3) consisted of nearly equal abundances of
black and scarlet oak. White oak (Quercus alba L.) was somewhat greater in importance on ELT 18 plots (side slopes
N/E aspects; deep soils; Miller, 1981) than on 17 plots (side slopes S/W aspects; deep soils). In contrast, shortleaf

18tb Central Hardwood Forest Conlerence 554

Table 2. Spearman correlation coefficients (with probability level below) between DCA ordination axis scores and
environmental variables (A) and between environmental variables and species Importance Values (B).
For species codes see Table 1.

(A) Axis 1 Axis 2

Slope position -0.29 -0.25

4 . 01 4.01

Aspect* -0.22 0.43

4.01 4.01

Aspect x -0.19 0.46

slope' 4.01 4 . 01

Slope (96) -0.02 -0.08 -0.10 -0.07 0.02 0.15 0.001 -0.07 0.09 0.12
0.59 0.04 0.01 0.10 0.69 <0.01 0.98 0.07 0.03 4.01

slope -0.13 -0.14 0.01 0.07 -0.12 -0.16 -0.13 -0.05 -0.05 -0.22
position 4.01 4.01 0.76 0.10 <0.01 <0.01 4.01 0.25 0.23 4.01

m' 0.01 0.03 0.29 0.15 0.19 0.26 0.03 0.10 -0.07 0.22
0.94 0.46 4.01 4.01 ~0.01 4.01 0.43 4.01 0.06 4.01

m x 0.01 0.05 0.33 0.15 0.2 1 0.28 0.7 0.12 -0.10 0.23
Slope' 0.90 0.24 4.01 4.01 4.01 4.01 0.07 4.01 0.01 4.01

QUeRALB QUERCOC QUERMAR QUERMUE QUERSTE QUERVEL SASSALB ULMUALA ULMURUB

Slope (5%) 0.10 0.12 0.15 -0.04 0.01 -0.26 -0.07 0.01 0.04 -0.04
4.01 4.01 4.01 0.32 0.81 4.01 0.06 0.99 0.34 0.36

kPped Wonnd trigonodcally (see text).

555 10th Central H a r d d Forest Conference

Table 3. Mean Tree Importance Values of MOFEP plots classified by Ecological Land Type (ELT). (n= number of
plots in ELT)

ELT '
1 5 6 7 11 15 17 18 19 20 21 22 23
n= 1 28 2 6 89 5 257 213 19 8 4 3 9
SPeCIES

ACERSAC*
CARYCOR
CARYGLA
C A R m
CARYTOM
CORNFLO
FRAXAME
JUGLNIG
JUNMR
wsssn
PINUECH
QUERALB
QueRCOC
QUERMAR
Q W U E
Q-TE
QUERVEL
SASSALB
ULMUALA
ULMURUB

*Species code: see Tib.1. FILT Code8 (Miller,1981): 1-Low floodplainslloamy and mixed alluvial soils; 5,bUpland waterwaysldeepcherty
soils; 7-Toe slopeddeep loamy soils; 11-Namnv ridgesldeepcherty soils; 15-Broad ridgesldeep loamy and dayey soils; 17-Side slopeslSouth and
West acpedsldeepsoils; 18-Side slopedNorth and East a q e d d e e p mils; 19-Side slopeslSouthand West aPpectslshallowsoils; 20-Side
slopWNorth and Past shallow soils; 21-23Side slopeshrery shallow limestone soils

pine was more important on ELT 17 plots. Other tree species on these plots generally had importance values greater
than 1 and less than 15 and included pignut, black and mockernut hickories (Carya glabra (Mill.) Sweet, C. texana
Buckl., C. tomentosa (Poir.) Nutt.), black gum (Nyssa sylvatica Marsh.) and post and blackjack oak. These hickory
species and black gum tended to be more important on ELT 18 plots, while blackjack and post oaks were more
important on ELT 17 plots. Plots of ELT 11 (narrow ridgeddeep cherty soils) were dominated by black oak with
smaller contributions by scarlet (Quercus coccinea Muenchh.), white and post oaks. Lower importance values were
associated with shortleaf pine, the three above-mentioned species of hickory, and black gum. Plots of ELT 19 (side
slopes S/W aspects; shallow soils) were characterized by the highest shortleaf pine and chinkapin oak (Quercus
muehlenbergiiEngelm.) importance values among all ELT classes, with similar contributions by scarlet, white, and
post oaks. Most important on plots characterized as "Upland Waterways" (ELT 5,6) was white oak, followed by
scarlet, post and black oak, and shortleaf pine and previously-mentioned species of hickory. Moreover, there was a
small, but notable, contribution on ELT 5 plots of mesic species such as black walnut, black gum, sugar maple
bitternut hickory, and slippery elm (Uimus rubra Miihl). Although the sample sizes of remaining ELT classes were
quite small, it is worth noting that white oak importance was quite high on ELT 20 (side slopes N/E aspects; shallow
soils) and those species that prefer soil high in calcium, such as eastern redcedar and chinkapin oak (F'owells 1965).
were important on ELT 21-23 plots.

Correlation of species importance values with environmental parameters (Table 2B) reinforced the qualitative
relationships noted above between ELT-based environmentalcharacterizations and species distribution and
abundance. Species associated with ELTs categorized as having South and West-facing slopes (shortleaf pine and

10th Central Hardwood Forest Conference 556

blackjack and post oaks) tended to be negatively correlated with the transformed aspect and aspect x slope 5%
variables. In contrast, species that were more important on ELTs associated with North- and East-facing slopes (white
oak, species of hickory, black gum, and black walnut) tended to be positively and significantlycorrelated with these
two parameters. Black oak correlated most highly with slope position, being most important on ridgetop positions
and showing little correlation with aspect. In contrast, more mesic-site species (for this region) such as sugar maple,
bitternut hickory, white ash, black gum, white oak, and winged (UZmus alata Michx.) and slippery elm were
associated with lower slope positions.

MOFEP plots show species composition and abundance typical of the generally xeric forests of the region (Braun
1972, Nelson 1985, Read 1952, Ware and others 1992). On ridgetop and upper slope forests a combination of black, '

scarlet, post and white oaks and shortleaf pine dominate, with relative importance shifting among black oak
(ridgetops), scarlet, black and post oaks and shortleaf pine (S- and W-facing slopes), and white oak and hickories (N-
and E-facing slopes). On lower slopes white oak tends to attain an importance similar to that associated with mesic
(North and East) aspects on upper slopes. On the relatively infrequent mesic sites associated with upland draws and
ephemeral streams, mesic species of the region such as sugar maple, bitternut hickory, black gum and black walnut,
along with more abundant white oak may be found.

According to Nelson's (1985) scheme, MOFEP forests would appear to variously fit the classifications of xeric to dry-
mesic forests on lirnestone/dolornite, chert and sandstone substrates. This conclusion is consistent with that of
Brookshire and Hauser (1993). MOFEP forests similarly fit into Braun's (1972) categorization of Oak-Pine and
southern division Oak-Hickory forests. There were some exceptions to Nelson's and Braun's descriptions, however.
Particularly noteworthy was the scarcity of northern red oak (Quercus ncbra L.) in MOFEP forests despite its
inclusion in forests of the region by both Nelson and Braun. Similarly, Ware and others (1992) identified northern
red oak as a common constituent of upland forests near the Current River in the same general area. Secondly, is the
abundance of scarlet oak in the MOFEP study sites and its restriction to the Oak-Chestnut Region by Braun (1972)
(she nowhere mentions it as occumng in Oak-Hickory forests). Scarlet oak also fails to appear in the species lists of
Ware and others (1992). It might be speculated that, given the morphological similarity of scarlet oak to northern red
and black oaks, variation in species identification might be responsible for these discrepancies.

Overstory-Understory Relationships

Figures 3-7 illustrate differences between overstory (>19.05 cm dbh) and understory k19.05 cm dbh) relative
densities of major species of common ELTs . Some trends were apparent: (1) In every ELT examined, there was
evidence of substantial depletion of understory populations of shortleaf pine and major red oak group species (scarlet
and black oak); (2) In three of five major ELT classes (11,17, 18, representing 92% of plots) white oak relative
density in the understory was significantly greater than that in the overstory; (3) A number of other species that
currently are not very important in the overstory show greater relative abundances in the understory [i.e., black gum,
mockernut hickory, black hickory and sassafras (Sassafras afbidwn (Nutt.) Nees)].

Overstory-understory relationships varied somewhat with ELT classification. The understory relative densities of red
oak group species and shortleaf pine tended to be greater in plots of ELTs that were more xeric in character
(especially ELT 11 and 17, Figs. 4 and 5) compared with more mesic ELTs (5 and 18 Figs. 3 and 6). However,
understory relative densities of these species were substantially reduced on ELT 19 plots, which also presumably was
quite xeric. As mentioned above, relative density of understory white oak was greater than that in the overstory in
three ELTs (11,17 and 18, Fig. 4,5 and 6). In the other two abundant ELTs (5 and 19) relative density of understory
white oak was less than that in the overstory, but because there were few black or scarlet oak and shortleaf pine plants
in the understory of forests of these ELTs, white oak was still the most common understory tree species. Other
species that tended to show increased relative density in the understory (black gum, black and mockernut hickory,
sassafras) showed little apparent relationship between relative abundance and ELT (and hence microenvironment).
Post oak was relatively less abundant in the understory than in the overstory in all ELTs, but the diiferences were
substantially less than those shown by red oak group species and shortleaf pine.

551 10th Central Hardwood Forest Conference

 * ELT 5 '""'"""
~ m < i r e m r ~

Species Species

Figure 3. Overstory and understory relative density Figure 4. Overstory and understory relative
for major species on ELT 5 plots. Asterisks above density for major species on ELT 11 plots.
each pair of histogram bars indicate highly significant Asterisks above each pair of histogram bars
(**@.01) and significant ( w . 0 5 )differences indicate highly significant ( * m . 0 1 ) and
between overstory and understory pairs. Bar pairs significant (*@.05) differencesbetween
with "ns" are not significantly different (p9.05). overstory and understory pairs. Bar pairs with
Species codes: Pe-shortleaf pine, Qc-scarlet oak; Qv- "ns" are not significantly different (p9.05).
black oak; Qs-post oak; Qa-white oak, Ns-black gum; Species codes: see legend for Fig. 3.
CT-black hickory; Ct-mockernut hickory; Sa-
sassafras..

ELT 1 7 1RD + 19 am dbh

0I D < lS am dbh
.- n-257

91 I S -
.?
&

-0
lo
1s-
. r
Ir" to- ns

-
s
o 1 nilnn-
Pe Qc QV Qs 00 Ns CT Ct So
Species Species

Figure 5. Overstory and understory relative density for Figure 6. Overstory and understory relative density
major species on ELT 17 plots. All bar pairs are for major species on ELT 18 plots. All bar pairs except
significantly different w.01). Species codes: see that marked with "ns" are significantly different
legend for Fig. 3. (p4.01). Species codes: see legend for Fig. 3.

10th Central Hardwood ForPst Conference 558

 -
.Z
CI
zo - *
-0 IS-
g 10- - nr nn
s -
O
PI
L 1 nmJ'A'
PC QV Or 00 Ns CT Ct So
Species

Figure 7. Overstory and understory relative density for

major species on ELT 19 plots. Asterisks above each pair
of histogram bars indicate highly significant (**pd.01)
and significant (*p4.05) differences between overstory
and understory pairs. Bar pairs with "ns" are not
significantly different (pM.05). Species codes: see legend
for Fig. 3.

Density-Size Class Distribution Relationships

%'-tests of distribution differences across size classes indicated that there were highly significant (psO.01) differences
among species in every ELT (data not shown). Density vs. size class plots were constructed for major species and
ELTs [see Fig. 8 for an example (ELT 17)]. In general, the results offered further confirmation of previously-noted
patterns of species dominance and potential replacement across ELTs.

However, several new distribution patterns were shown in these plots and not elsewhere: (1) In all ELTs white oak
abundance was greatest in the 1.5-4.5 in (3.8-1 1.4 cm) dbh size class, decreasing in all larger and the one smaller size
class; (2) Most other abundant species in the understory (black gum, sassafras, black hickory, mockernut hickory)
had greatest density in the smallest size class; (3) Some species that currently are abundant in the overstory (black,
scarlet, and post oak and shortleaf pine) showed shallow negative slopes; (4) Density of red oak group species and
shortleaf pine were much reduced in smaller size classes compared with those of white oak and other understory
species, supporting the general trends and conclusions shown by overstory-understoryanalysis (see above).
However, red oak group species are present in densities > 50ha even in smaller size classes in c d n ELTs,
particularly those that are more xeric (e.g., 17, Fig. 8). Density of shortleaf pine is everywhere less than 25lha; (5) The
differences between ELT 17 and 18 in relative density of red oak group species and shortleaf pine vs. white oak were
attributable to differences in absolute abundance in the former group. Absolute density of white oak in the understory
was quite similar on both ELTs (data not shown; see also Brookshire and Hauser 1993).

559 10th Central H a r d w d Forest Conference

 <3.87.6 15.2 22.8 30.5 38.1 45.7 53.3 61.0 68.6+
Diameter class midpoint (cm dbh)
Figure 8. Mean density vs. size class relationships for important species on ELT 17 plots. Species codes:
see legend for Fig. 3.

These data are consistent with trends we have seen elsewhere in the Ozarks (Jenkins and Pallardy, 1993). There is
strong evidence in the MOFEP data set and from other research in the Missouri Ozarks that white oak has the potential
to increase in importance thl-oughoutthe region. The ecological basis of this potential change may be ascribable to
white oak's greater longevity and shade tolerance (Dougherty and others 1980, Fowells 1965 ). The size class
distributions further indicate that white oak apparently has experienced a "pulse" of recruitment, with the greatest
abundance of this species found in the next-to-smallest size class. Other species, particularly black gum, some of the
hickories, and sassafras show great abundance in the smallest size class. One might speculate that observers of long-
term f m t succession in the Ozark region might see these species becoming more important, but it is debatable
whether they (except perhaps for the hickories) have canopy potential given the xeric environment of the Ozarks and
the greater microclimatic stress associated with ascent to an unprotected canopy position (Nigh and others 1985,
Pallardy and others 1988). On the other hand, the current position of white oak as a canopy dominant throughout
much of the MOFW study area indicates that its short-term increase in abundance is nearly certain.

Future trends in abundance of scarlet and black oak in MOFEP site forests may differ somewhat from those of
shortleaf pine. In the former pair there is still sufficient regeneration to anticipate a continuing, if less prominent,
contribution of both red oak species to future forest overstories. This is especially true on xeric ELT 11, 17, and 19
sites. In contrast, seedlings and saplings of shortleaf pine are nowhere abundant and one might predict a uniform
decline in the abundance of this species through time in canopies of undisturbed forests of the M O W project.

ACKNOWLEDGMENTS

Funding of this work was provided by the Missouri Department of Conservation and is gratefully acknowledged.

~ WCentrsl
I Hardwood Forest Conference 560
Beers, T. W., P. E. Dress, and L. C. Wensel. 1966. Aspect transformation in site productivity research. J. For. 64:
691- 692.

Braun, E. L. (1972). Deciduous Forests of Eastern North America. Hafner Publ. Co., New York.

Brookshire, B., and C. Hauser. 1993. The Missouri Forest Ecosystem Project. Pp. 289-307. In;A. R. Gillespie, G. R.
Parker and P. E. Pope, eds. Proc. Ninth Central Hdwd. For. Conf., Purdue University. U.S.D.A. Forest
Service, Gen. Tech. Rep. NC-16 1.

Dollar, K. E., S. G. Pallardy, and H. E. Garrett. 1992. Composition and environment of floodplain forests of northern
Missouri. Can. J. For. Res. 22: 1343-1350.

Dougherty, P. M., G. S. Cox, and T. M. Hinckley. 1980, A summary of information relating to the ecophysiology of
white oak (Quercur alba L.). Pp. 116-135. In: H. E. Garrett and G. S. Cox, eds. Proc. Third Central Hdwd.
F a . Conf., U. of Missouri, Columbia, MO.

Fowells, H. A. (1965). Silvics of the Forest Trees of the U.S. U.S.D.A. Forest Service, Agric. Hdbk. No. 271.

Frank, E. C., and R. Lee. 1966. Potential solar beam irradiation and slopes for 30 to 50 degrees latitude. U.S.D.A.
Forest Service, Rocky Mountain For. and Range Expt. Sta., Res. Pap. 18.

Harcombe, P. A. 1987. Tree life tables. Bioscience 37: 557-568.

Harcombe, P. A., and P. L. Marks. 1978. Tree diameter distributions and replacement processes in southeast Texas
forests. For. Sci. 24: 153-166.

Hill, M. 0. 1979. TWINSPAN--A FORTRAN program for arranging multivariate data in an ordered two-way table
by classification of the individuals and attributes. Cornell University, Ithaca, New York.

Hill, M. O., and H. G. Gauch. 1980. Detrended correspondence analysis: an improved ordination technique. Vegetatio
42: 47-58.

Jenkins. M. A., and S. G. Pallardy. 1993. A comparison of forest dynamics at two sites in the southeastern Ozark
Mountains of Missouri. Pp. 327-341. Ip;A. R. Gillespie, G. R. Parker and P. E. Pope, eds. Proc. Ninth Central
Hdwd. For. Conf., Purdue University. U.SJ3.A. Forest Service, Gen. Tech. Rep. NC-161.

Johnson, P. S. 1992. Oak overstory reproduction relations in two xeric ecosystems in Michigan. For. Ecol. Mgmt. 48:
233-248.

McCarthy, B. C., and W. A. Wistendahl. 1988. Hickory (Carya spp.) distribution and replacement in a second-growth
oak hickory forest of southeastern Ohio. Amer. Midl. Nat. 119: 156-164.

Miller, M. R. 1981. Ecological Land Classification Terrestrial Subsystem: Mark Twain National Forest. A Basic
Inventory System for Planning and Management. U. S. D. A. Forest Serv., Resources Planning and
Management, Rolla, Missouri.

Nelson, P. W. 1985. The Terrestrial Natural Communities of Missouri. Missouri Departments of Natural Resources
and Conservation, Jefferson City, Missouri.

561 10th Central Hardwood Forest Conference

Nigh, T. A., S. G. Pallardy, and H. E. Garrett. 1985. Sugar maple-environment relationships in the River Hills and
central Ozark Mountains of Missouri. Amer. Midl. Nat. 114: 235-251.

Pallardy, S. G., T. A. Nigh, H. E. Garrett. 1988. Changes in forest composition in central Missouri: 1968-1982. Amer.
Midl. Nat. 120: 380-390.

Read, R. A. 1952. Tree species occurrence as influenced by geology and soil on an Ozark north slope. Ecology 33:
239-246.

SAS Institute, Inc. 1988. User's Guide to the Statistical Analysis System: Statistics. SAS Institute, Cary, NC.

Sokal, R. R, and F. J. Rohlf. 1981. Biometry. Second edition. Freeman, San Francisco.

Ware, S., P. L. Redfearn., G. L. Pyrah, and W. R. Weber. 1992. Soil pH, topography and forest vegetation in the
central Ozarks. Amer. Midl. Nat. 128: 40-52.

Zimmennan, M., and W. L. Wagner. 1979. A description of the woody vegetation of oak-hickory forest in the
northern Ozark highlands. Bull. Torrey Bot. Club 106:117-122.

loth Central Hardwood Forest Conference 562