This document is Philip Hofmeyer's doctoral thesis submitted to the University of Maine in 2008. It examines the ecology and silviculture of northern white-cedar in Maine. The thesis contains three chapters: 1) A literature review synthesizing northern white-cedar research findings across its northeastern North American range; 2) An analysis of the influence of soil site class on northern white-cedar growth and decay patterns using field data from 60 sites in Maine; and 3) Development of models to predict northern white-cedar leaf area and stemwood growth efficiency.

1. ECOLOGY AND SILVICULTURE OF NORTHERN WHITE-CEDAR
(THUJA OCCIDENTALIS L.) IN MAINE.
By
Philip V. Hofmeyer
A.A.S. State University of New York at Morrisville, 1999
B.S. State University of New York College of Environmental Science and Forestry, 2001
M.S. State University of New York College of Environmental Science and Forestry, 2004
A THESIS
Submitted in Partial Fulfillment of the
Requirements for the Degree of
Doctor of Philosophy
(in Forest Resources)
The Graduate School
University of Maine
May 2008
Advisory Committee:
Laura S. Kenefic, Research Forester and Silviculturist, U.S. Forest Service,
Northern Research Station and Forest Resources Faculty Associate, Co-advisor
Robert S. Seymour, Curtis Hutchins Professor of Forest Resources, Co-advisor
John C. Brissette, Research Forester and Project Leader, U.S. Forest Service,
Northern Research Station
Ivan J. Fernandez, Professor of Soil Science
William H. Livingston, Associate Professor of Forest Resources

2. LIBRARY RIGHTS STATEMENT
In presenting this thesis in partial fulfillment of the requirements for an advanced
degree at The University of Maine, I agree that the Library shall make it freely available
for inspection. I further agree that permission for “fair use” copying of this thesis for
scholarly purposes may be granted by the Librarian. It is understood that any copying or
publication of this thesis for financial gain shall not be allowed without my written
permission.
Signature
Date:

3. ECOLOGY AND SILVICULTURE OF NORTHERN WHITE-CEDAR
(THUJA OCCIDENTALIS L.) IN MAINE.
By Philip V. Hofmeyer
Thesis Co-advisors: Dr. Laura S. Kenefic and Dr. Robert S. Seymour
An Abstract of the Thesis Presented
in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
(in Forest Resources)
May, 2008
Northern white-cedar (Thuja occidentalis L.) management has been hindered
throughout is native range in part because of a lack of fundamental ecology and
silviculture research. Efforts to tend and regenerate northern white-cedar stands
frequently yield inconsistent and unpredictable results due to disagreement regarding of
its ecology. In the present study, two breast-height cores were extracted from 625
outwardly sound sample trees at 60 sites in northern Maine. Northern white-cedar annual
basal area growth predicted from breast height sapwood area was compared to that red
spruce (Picea rubens Sarg.) and balsam fir (Abies balsamea (L.) Mill.) along site class
and light exposure gradients. A subsample of 25 sound northern white-cedar trees was
stem-analyzed to develop allometric leaf area equations and test for differences in growth
efficiency by site and light exposure. An additional 59 sound northern white-cedar trees
were stem-analyzed to reconstruct early height and diameter development. Results
suggest that northern white-cedar growth is not strongly affected by site class or light

4. exposure class. Central decay from heart rot fungi occurred in nearly 80% of the northern
white-cedar trees sampled. Incidence of decay and proportion of basal area centrally
decayed increased as soil drainage improved. Projected leaf area and crown foliage mass
were estimated with a nonlinear basal area and live crown ratio model. Volume increment
per unit leaf area was modeled with a two-parameter nonlinear power function. Growth
efficiency was not strongly influenced by site class, canopy position, breast height age, or
presence of central decay. Early stem data suggest that many of the sound cedar trees
sampled had a history as advance regeneration. Early height and diameter growth were
slow, though most trees had a period of release in their ring chronology coinciding with
known spruce budworm (Choristoneura fumiferana) epidemics. Managers are
recommended to favor individuals with large crowns as residuals in partial harvests
regardless of site class. Northern white-cedar could likely be managed successfully with
uneven-aged silviculture or variants of the shelterwood system. Caution should be taken
to avoid residual stand damage to seedlings and saplings during harvesting operations.

5. ii
ACKNOWLEDGEMENTS
Scientific undertakings are commonly built upon the support, guidance, and
intellectual input from many others. This study was surely no exception. There are many
people to whom I owe gratitude far beyond what can be written on paper. Dr. Laura
Kenefic and Dr. Robert Seymour have been incredible advisors to me. Laura brought me
onto the project and has truly allowed me to learn the value of cooperative research
among scientists with similar objectives. Bob’s uncanny knack for quickly identifying
quantitative hiccups has saved me on many occasions. I would also like to thank Dr. Bill
Livingston, Dr. Ivan Fernandez, and Dr. John Brissette for their guidance as members of
my committee.
For providing financial support and study sites for this project I would like to
thank the University of Maine’s Cooperative Forestry Research Unit and all of its
members as well as the School of Forest Resources, and Maibec Industries, Inc. In
particular, I would like to thank Charles Tardif for his tireless energy dedicated to
understanding the ecology, siliviculture, and processing of northern white-cedar.
I would like to thank Zachary Bergen for providing a summer of stories, bug
swatting, and destructive sampling in the North Maine Woods. I would like to thank
Kersi Contractor for a semester spent separating dried foliage and reading over 3000
cedar chronologies with me. Ken Laustsen with the Maine Forest Service went over and
above any expectations I might have had with regards to providing Forest Inventory and
Analysis data for this study. The support and assistance from these gentlemen is greatly
appreciated.

6. iii
Several fellow students at the University of Maine were instrumental in providing
technical and moral support throughout my time spent working on this project. I would
like to thank are Jamie Weaver, Justin Waskeiwitcz, David Ray, and Spencer Meyer. I
would also like to thank Catherine Larouche at the University of Laval for her support
and thoughts on cedar regeneration in Maine and Quebec.
Finally, I owe gratitude and love to my wife, Jessica, for her undying support of
me throughout graduate school. She provided an outlet for my ranting, the kindling that
lit the fire under me to stay focused, and the sanctuary of a peaceful home. This project
would never have been completed without her encouragement.

7. iv
TABLE OF CONTENTS
ACKNOWLEDGEMENTS............................................................................................ ii
LIST OF TABLES........................................................................................................vii
LIST OF FIGURES ....................................................................................................... ix
Chapter 1: NORTHERN WHITE-CEDAR ECOLOGY AND SILVICULTURE
IN THE NORTHEASTERN UNITED STATES AND SOUTHEASTERN
CANADA: A SYNTHESIS OF KNOWLEDGE BY REGION......................................... 1
1.1 ABSTRACT.................................................................................................... 1
1.2 INTRODUCTION .......................................................................................... 2
1.3 REGIONAL FINDINGS ................................................................................ 4
1.3.1 Ontario ........................................................................................................ 4
1.3.2 Quebec ........................................................................................................ 5
1.3.3 Lake States.................................................................................................. 6
1.3.4 The Northeast............................................................................................ 11
1.4 ECOTYPIC VARIATION............................................................................ 12
1.5 SUMMARY.................................................................................................. 14
Chapter 2: INFLUENCE OF SOIL SITE CLASS ON GROWTH AND DECAY
OF NORTHERN WHITE-CEDAR.................................................................................. 16
2.1 ABSTRACT.................................................................................................. 16
2.2 INTRODUCTION ........................................................................................ 17
2.3 METHODS ................................................................................................... 19
2.3.1 Site Description......................................................................................... 19
2.3.2 Field Data Collection................................................................................ 21
2.3.3 Data Analysis............................................................................................ 24
2.4 RESULTS ..................................................................................................... 26
2.4.1 Basal Area Growth.................................................................................... 26
2.4.2 Decay ........................................................................................................ 27
2.4.3 Site Index .................................................................................................. 27

8. v
2.5 DISCUSSION............................................................................................... 32
2.5.1 Basal Area Growth.................................................................................... 32
2.5.2 Stem Decay............................................................................................... 34
2.5.3 Site Index .................................................................................................. 37
2.6 CONCLUSIONS........................................................................................... 39
Chapter 3: LEAF AREA PREDICTION MODELS AND STEMWOOD
GROWTH EFFICIENCY OF THUJA OCCIDENTALIS L. IN MAINE ....................... 41
3.1 ABSTRACT.................................................................................................. 41
3.2 INTRODUCTION ........................................................................................ 42
3.3 METHODS ................................................................................................... 44
3.3.1 Field Methods ........................................................................................... 44
3.3.2 Laboratory Procedure................................................................................ 48
3.3.3 Statistical analysis..................................................................................... 50
3.4 RESULTS ..................................................................................................... 55
3.4.1 Branch Leaf Area and Foliage Mass......................................................... 55
3.4.2 Projected Leaf Area and Crown Foliage Mass ......................................... 58
3.4.3 Stem Volume ............................................................................................ 60
3.4.4 Volume Increment and Growth Efficiency............................................... 60
3.5 DISCUSSION............................................................................................... 65
3.5.1 Leaf Area Prediction................................................................................. 65
3.5.2 Volume Increment and Growth Efficiency............................................... 66
3.6 CONCLUSIONS........................................................................................... 70
Chapter 4: HISTORICAL EARLY STEM DEVELOPMENT OF NORTHERN
WHITE-CEDAR IN MAINE ........................................................................................... 72
4.1 ABSTRACT.................................................................................................. 72
4.2 INTRODUCTION ........................................................................................ 73
4.3 METHODS ................................................................................................... 77
4.4 RESULTS ..................................................................................................... 81

9. vi
4.5 DISCUSSION............................................................................................... 86
4.6 CONCLUSIONS........................................................................................... 92
LITERATURE CITED................................................................................................. 93
APPENDICES ................................................................................................................ 105
Appendix A: SITE INDEX RING CHRONOLOGIES.............................................. 106
Appendix B: FIT STATISTICS FOR PROJECTED LEAF AREA AND
CROWN FOLIAGE MASS REGRESSION MODELS............................................. 125
Appendix C: STEM-ANALYZED TREE RING BREAST HEIGHT
CHRONOLOGIES ..................................................................................................... 127
Appendix D: MEAN ANNUAL INCREMENT STEM PROFILES.......................... 141
BIOGRAPHY OF THE AUTHOR............................................................................. 146

10. vii
LIST OF TABLES
Table 2.1 Briggs (1994) site class descriptions……………………………………....23
Table 2.2 Mean height, diameter, and breast height age of sample trees…………….23
Table 2.3 Number of sample trees by site and light exposure classes………………..24
Table 2.4 Northern white-cedar sapwood area per unit basal area (SA:BA)
by light exposure class…………………………………………………..…29
Table 2.5 Mean basal area growth as a function of site class and light exposure
class with a significant sapwood area covariate……………………….......29
Table 2.6 Mean incidence of sample trees decayed by site class (standard
errors in parentheses)……………………………..………………………..30
Table 2.7 Site index (m) at breast height age 50 for balsam fir, northern white-
cedar and red spruce by site class determined from trees without
radial growth suppression (standard errors in parentheses) in light
exposure classes 3 through 5……………………………………..………...31
Table 3.1 Attributes of 25 destructively sampled northern white-cedar trees……..…47
Table 3.2 Mean specific leaf area, branch foliage mass, and branch leaf area
for the lower, mid, and top crown sections of 25 northern white-
cedar sample trees…………………………………………..…………...…50
Table 3.3 Projected leaf area (PLA) and crown foliage mass (CFM) models
fit to 25 northern white-cedar sample trees (only PLA shown as
dependent variable)………………………………………………………...52

11. viii
Table 3.4 Specific leaf area (cm2
/g) with respect to location within the crown,
among site classes, and among light exposure classes………………….…56
Table 3.5 Model evaluation for branch leaf area (cm2
) and branch foliage
mass (g)………………………………………………………………….....56
Table 3.6 Best-fit area inside bark (AIB), basal area outside bark (BA), crown
length (CL), and sapwood area (SA) model for estimating projected
leaf area (PLA) and crown foliage mass (CFM), ranked by FI values…….58
Table 3.7 Parameters and fit statistics of the nonlinear regressions of VINC
on PLA and CFM for stem-analyzed and cored trees……………………...61
Table 3.8 ANOVA results for growth efficiency of 25 destructively sampled
northern white-cedar trees by site and light exposure class………………..63
Table 3.9 ANOVA results for growth efficiency of 296 cored northern
white-cedar trees by site and light exposure class…………………………65
Table 4.1 Stand-level characteristics for 21 stem-analyzed northern white-
cedar trees………………………..………………………………………...79
Table 4.2 Mean annual height increment (m) of 80 stem-analyzed northern
white-cedar trees…………………………………………………………...82
Table 4.3 Number of years required to reach a given inside bark diameter at
a given height for 80 stem-analyzed northern white-cedar trees…………..82
Table B.1 Fit statistics for PLA models………..…………………………………….125
Table B.2 Fit statistics for CFM models………..……………………………………126

12. ix
LIST OF FIGURES
Figure 1.1 Distribution of northern white-cedar………………………..…………….3
Figure 2.1 Study site locations throughout central and northern Maine………..…...20
Figure 2.2 Basal area (BA) growth as a function of sapwood area (SA) by
species……………………………………………………………………28
Figure 2.3 Proportion of sample tree basal area decayed by site class in
northern white-cedar, balsam fir, and red spruce………………………..30
Figure 3.1 Site locations of the 25 destructively sampled northern white-
cedar trees………………………………………………………………..46
Figure 3.2 Observed and predicted branch leaf area as a function of branch
diameter (A) and relative distance of the branch into the crown
(B) with model [2a]……………………………………………………...57
Figure 3.3 Branch leaf area (A) and foliage mass (B) as a function of relative
distance into the crown and branch diameter as predicted by
model [2a]………………………………………………………………..57
Figure 3.4 Projected leaf area and crown foliage mass values calculated by
branch summation for 25 stem-analyzed cedar trees and estimated
from their respective tree core data with model [BA 4]…………………59
Figure 3.5 Observed stemwood volume analyzed in WinDendro with the
estimated volume from Honer’s (1967) equation with refit
parameter coefficients…………………………………………………...61

13. x
Figure 3.6 Comparison of monotonic decreasing regression model [4]
relating annual stemwood volume increment to projected leaf area
for 25 stem-analyzed and 256 cored northern white-cedar trees……...…62
Figure 3.7 Growth efficiency as a function of projected leaf area (A) and
crown foliage mass (B) by light exposure class in 296 cored
northern white-cedar trees……………………………….………...…….64
Figure 3.8 Growth efficiency as a function of breast height age……………………68
Figure 4.1 Timberland area (A) and growing stock volume (B) of northern
white-cedar by stand size class in Maine………………………………..75
Figure 4.2 Site locations of the destructively sampled northern white-cedar
stems from Maibec Industries (◊) and the growth efficiency
study (□)…………................................................................................…78
Figure 4.3 Typical pattern of suppression followed by a release and relatively
constant radial growth in the stump height disc (top) and no signs
of a suppressed core at the mid height disc (bottom)……………………83
Figure 4.4 Stem profile of sample tree M 42. Each line represents one year
of height and diameter growth……………….…….……………………84
Figure 4.5 Breast height ring chronologies of four northern white-cedar trees
in Maine…………………..…………………………………………….85
Figure 4.6 Mean annual area and radial increment for Tree LA 6 consistent
with Pressler’s Law……………………………………………………...90
Figure A.1 Ring chronologies of the sample trees selected to quantify site
index…………………………………………………………………....106

14. xi
Figure C.1 Breast height ring chronology of 80 northern white-cedar trees
from central and northern Maine stem-analyzed for early stem
development patterns…………………………………………………...127
Figure D.1 Mean annual area increment as a function of disc height in 25
stem-analyzed northern white-cedar trees……………………………...141

15. 1
Chapter 1:
NORTHERN WHITE-CEDAR ECOLOGY AND SILVICULTURE IN THE
NORTHEASTERN UNITED STATES AND SOUTHEASTERN CANADA: A
SYNTHESIS OF KNOWLEDGE BY REGION
1.1 ABSTRACT
Sustainability of the northern white-cedar (Thuja occidentalis L.) resource is a
concern in many regions throughout its range because of regeneration failures, difficulty
recruiting seedlings into sapling and pole classes, and harvesting levels that exceed
growth. Management confusion has resulted from the scarcity of research on northern
white-cedar ecology and silviculture, as well as apparent regional differences in findings.
This paper synthesizes recent and historical northern white-cedar literature, with findings
presented by region. Though a number of past studies have produced contradictory
findings, some generalizations of use to the practitioner can be made; northern white-
cedar is small-stature, decay-prone except on cliff sites, and found in both early- and late-
successional stands. Northern white-cedar appears to be a highly variable species that can
adapt to a wide range of environmental stresses. Regional inconsistencies in cedar
ecology and silviculture limit widespread applicability of many studies and support the
development of local guidelines.

16. 2
1.2 INTRODUCTION
Northern white-cedar (Thuja occidentalis L., hereafter abbreviated “NWC”), is
common throughout southeastern Canada and the northeastern United States; its range
extends from Minnesota to Nova Scotia (Figure 1). Local populations can also be found
south along the Appalachian Mountains from Pennsylvania to Tennessee (Johnston
1990). Despite its abundance, NWC is arguably the least studied commercially valuable
tree species in its region (Scott and Murphy 1987). Studies of NWC ecology and
silviculture are often limited in scope and geographical range (Hofmeyer et al. 2007);
very little is known about NWC growth and management in the Northeastern United
States. Many useful papers about NWC were published in conference and workshop
proceedings, or as university or government reports, as early as the 1910s. Such
literature is not readily accessible to the practitioner.
NWC stands in many regions are impacted by browsing, harvesting, competition
from associated tree and shrub species, and recruitment failures. Recent U.S. Forest
Service Forest Inventory and Analysis (FIA) data from Maine, for example, suggest that
there has been an annual negative net change of approximately 245,000 m3
of NWC
growing stock since 1995 (McWilliams et al. 2005). This was primarily attributed to a
lack of ingrowth, recruitment of poletimber to sawtimber without replacement, increases
in cull volume, and harvest levels that exceeded net growth. In the Lake States, NWC is a
common deeryard species because it provides critical winter habitat for white-tailed deer
(Odocoileus virginianus). Regeneration and maintenance of stand structure has become
problematic in many deeryards, leading to concerns about the sustainability of these
stands (e.g. Miller et al. 1990, Van Deelen et al. 1996, Van Deelen 1999). With

17. 3
Figure 1.1. Distribution of northern white-cedar (Thuja occidentalis L.) (Little 1971).
suggestions that recent NWC harvesting may not be sustainable throughout portions of its
range, shortcomings in the NWC literature are strongly felt by forest managers.
Our objective was to synthesize NWC research findings by region, and to outline
current knowledge relevant to NWC management. Information for this review was
collected in a year-long literature search; a comprehensive list of NWC literature
(English-language only, published before April 2007) relevant to forestry can be found in
Northern White-Cedar: An Annotated Bibliography (Hofmeyer et al. 2007).

18. 4
1.3 REGIONAL FINDINGS
1.3.1 Ontario
The majority of the NWC literature in Ontario comes from the Cliff Ecology
Research Group (CERG) at the University of Guelph, investigating limestone cliffs along
the Niagara Escarpment. These NWC-dominated forests are perhaps the most extensive
old-growth forests in eastern North America (Larson and Kelly 1991). The Niagara
Escarpment forests are considered to be free from large-scale disturbance, though rockfall
is a common small-scale disturbance (Kelly and Larson 1997). Portions of root systems
are often damaged by rockfall disturbances. Links between damaged portions of root and
shoot systems led researchers to discover that some NWC trees have a radially sectored
architecture that allows the tree to continue growing when portions of the roots, shoots,
and cambium die (Larson et al. 1993, Larson et al. 1994). Investigations have
demonstrated that some NWC trees possess stem stripping (alternating vertical bands of
living and dead wood); researchers hypothesize that this results from cavitation events in
old trees and allows for stress tolerance in harsh environments (Matthes-Sears et al.
2002). NWC trees on these sites have been compared to bristlecone pine (Pinus longaeva
D.K. Bailey) due to their advanced ages (some exceeding 1000 years), distorted
architecture, slow radial growth rates (<0.1mm/yr), and cambial mortality in living
specimens (Kelly et al. 1992). Though the Niagara Escarpment is a harsh site, there is
apparently adequate moisture, nutrients, and mycorrhizal colonization for NWC growth
(Matthes-Sears et al. 1992, 1995). As such, there has been extensive work by the CERG
to compare and contrast cliff NWC to non-cliff NWC.

19. 5
Cliff-dwelling NWC trees have a higher specific gravity, crushing strength, and
modulus of elasticity and rupture than non-cliff NWC (Larson 2001); many of these
properties have been linked to slow growth rates a high proportion of lignin-rich
latewood. Larson (2001) also reported a typical lifespan of only 80 years (maximum of
400 years) for non-cliff NWC, while NWC ages on cliffs have exceeded 1030 years
(Kelly et al. 1992). Aside from age, growth rate, and strength properties, cliff NWC trees
are considerably different from other NWC populations in that they rarely are afflicted
with central decay. Resistance to central decay, long lifespan, and stability of wood
structure after tree death has led to extensive use of cliff NWC in dendroclimatology and
dendroecology research (e.g. Kelly et al. 1992, 1994, Buckley et al. 2004).
1.3.2 Quebec
Research in northwestern Quebec has confirmed that NWC is not solely a short-
lived tree restricted to swamp sites, as many practitioners formerly believed. Specimens
exceeding 800 years of age have been discovered on xeric sites and used for
dendroclimatology (Archambault and Bergeron 1992). NWC trees from hydric sites have
also been used in northwestern Quebec in dendroclimatological work (Tardif and
Bergeron 1997). In both northwestern Quebec and southern Ontario, radial growth has
been correlated with the previous summer temperature and moisture; reduced radial
growth was observed in years that followed a hot, dry summer (Kelly et al. 1994, Tardif
and Bergeron 1997).
Land use history of southern Quebec is similar to much of the northeastern United
States in that land clearing and farm abandonment have had great impact on present day
forest communities (de Blois and Bouchard 1995). They found that NWC typically exists

20. 6
in dense, pure communities that resist colonization by competing species. When land use
practices modify soil and vegetation, NWC can invade more mesic sites. NWC
colonization occurred on 95 percent of abandoned pasture lands in a study in southern
Quebec; 68 percent of those were mesic sites (de Blois and Bouchard 1995). Sixty to 80
percent of NWC trees on some abandoned pasture lands in that region were of vegetative
origin, with low genetic diversity even among individuals in mixed-species associations
(Lamy et al. 1999). NWC trees in those stands had low levels of outcrossing, high levels
of self-fertilization, and high rates of vegetative reproduction.
NWC regeneration requirements have historically been difficult to determine and
even more difficult to manipulate. Research on regeneration requirements in Quebec
suggested that seedlings are susceptible to desiccation in highly disturbed stands, but that
established large seedlings and small saplings increase in height growth proportional to
increased light levels (LaRouche et al. 2006). NWC herbivory by white-tailed deer was
shown to be detrimental only at higher population levels in this region (Larouche et al.
2007).
1.3.3 Lake States
Most of the research on NWC ecology and silviculture comes from Michigan,
Minnesota, and Wisconsin. In 1967, Caulkins foretold that “…cedar will probably
receive more attention and study in the future than some of its commercially valuable
counterparts. This is because it is by far the most important food and winter cover for
North America’s number one big game animal, the white-tailed deer.” Though NWC has
been largely neglected in research throughout its range, there are many studies from the
Lake States on deeryard management, regeneration, and wildlife–cedar interactions.

21. 7
NWC stands are commonly used by white-tailed deer for overwintering; in
regions with significant snowfall, browse opportunities are often limited to tree shoots.
NWC has been found to be more palatable than several associated species, including
aspen (Populus spp.), jack pine (Pinus banksiana Lamb.) and balsam fir [Abies balsamea
(L.) Mill.] (Ullrey et al. 1964, 1967, 1968). Though deer often lose body mass on a
single-species diet, NWC can support them through harsh winters if there is at least 2 kg
of cedar browse per animal per day (Aldous 1941). In addition to browse opportunities,
deer frequently congregate in deeryards when there is more than 30 cm of snow to benefit
from reduced snow cover and wind, communal trails, more stable, warmer temperatures,
and predator avoidance (Sabine et al. 2001). Many of the challenges faced regarding
managing NWC deeryards are related to the use patterns of deer herds.
Historically, NWC swamps in northern Michigan were multi-storied stands that
had plentiful browse opportunities for deer, but sparse winter cover (Verme 1965). Many
NWC communities in the Lake States today originated after clearcutting in the early
1900s, a time period with relatively low deer populations (Heitzman et al. 1997, 1999).
These even-aged stands offer more protection but less available browse, though deer
populations in the Lake States are much higher than they have been in the past (Heitzman
et al. 1997). An increase in deer populations, coupled with NWC’s palatability, has led to
difficulties regenerating NWC stands in that region.
Multiple studies have documented slow NWC seedling and sapling height growth
rates (e.g. Johnston 1990, Heitzman et al. 1997, Davis et al. 1998). White-tailed deer
have been observed in wintering areas until April, preferentially browsing on NWC (Van
Deelen et al. 1999). Because of this browse preference and slow seedling growth rate, up

22. 8
to 40 years may be required for seedlings to grow out of deer browsing height (Van
Deelen et al. 1999). Herbivory pressure from white-tailed deer and snowshoe hare has led
researchers to investigate the use of animal exclosures during the regeneration period
(Miller 1990); more than 75 percent of NWC seedlings and saplings outside exclosures in
some areas have been excessively browsed within three years (Cornett et al. 2000).
In the Lake States, much of the NWC research has focused on overcoming browse
pressure on stand regeneration. One of the most extensive NWC regeneration studies
came from the Michigan Department of Conservation Game Division (Nelson 1951); this
study outlined several principles that continue to guide management today. Nelson (1951)
found that soil pH below 4.0 negatively affected germination, soil pH below 6.0
negatively affected seedling density, browsing and desiccation were the most common
causes of seedling mortality, vegetative reproduction was common (primarily layering),
and growth and development in the seedling stage is slower than competing species such
as balsam fir and tolerant hardwoods.
NWC regeneration is complex due in part to the difficulties in determining its
successional niche and shade tolerance. On sand dune sites in the Lake States, NWC can
act as a colonizing pioneer species that is replaced by shade-tolerant hardwoods (Scott
and Murphy 1987). Though NWC’s shade tolerance has been noted to range from shade-
tolerant (slightly less than balsam fir) to moderately intolerant (slightly less than white
pine) (Curtis 1946), it is typically classified as a shade-tolerant species that requires
significant disturbance to replace itself (Scott and Murphy 1986). In attempts to mimic
this regeneration requirement, strip clearcutting and strip shelterwood regeneration
methods have been recommended silvicultural treatments (Johnston 1977). These

23. 9
prescriptions tended to have little success in NWC regeneration due to the detrimental
effects of browse, logging slash, and desiccation (Thornton 1957b, Heitzman et al. 1999).
Research suggests that NWC seedlings can survive on nurse logs in later stages of
decay that hold moisture through the dry portions of the summer (Caulkins 1967); nurse
logs with an associated bryophyte mat are highly desirable for NWC seedlings
(Holcombe 1976). Seedling densities have also been positively correlated with the
proportion of the forest floor in hummocks on lowland pit and mound topography sites.
In Michigan’s Upper Peninsula, hardwood brush dominated the understory of NWC
stands with less than 70 percent of the ground area in hummocks (Chimner and Hart
1996); NWC regeneration was observed to be successful if the ground area was greater
than 70 percent hummocks. In nearly all cases, NWC seedling survival increases with
adequate moisture.
Many sites with NWC in the overstory are regenerating to competing species due
to the difficulties in recruiting NWC seedlings to the sapling stage. Balsam fir has been
shown to quickly overtop NWC seedlings on sites without a large component of downed
woody material (Cornett at al. 1997); this may be because fir’s larger seed size and
quickly developing root system enables it to better withstand drought periods. In addition,
Davis et al. (1998) found that while high numbers of NWC seedlings can be recruited
after low intensity ground fires, plots without deer exclosures had no NWC seedlings
after 10 years. Balsam fir was common on those plots, likely due to preferential
browsing of NWC and avoidance of balsam fir by white-tailed deer and snowshoe hare.
In some stands where NWC is a component of a mixed-species association, shade-

24. 10
tolerant hardwoods are expected to dominate in the future due to slow NWC seedling
growth (e.g. Thornton 1957a, Scott and Murphy 1987, Cornett et al. 2000).
Volume and yield tables were constructed for the Lake States based on a sample
of 227 stems distributed throughout Michigan, Minnesota and Wisconsin (Gevorkiantz
and Duerr 1939). These tables suggest that NWC trees taper abruptly from the stump
upward, commonly have a site index of 14 to 18 meters at 50 years at stump height, and
can form dense stands with high yields.
Early thinning trials on swamp sites suggested that reducing stand basal area
increased quality and vigor of the residual trees (Roe 1947), though response to thinning
was better on lowland sites with moving ground water than on sites with stagnant
groundwater. Many of the lowland stands were previously clearcut, diameter-limit cut
and/or otherwise selectively harvested, though Thornton (1957a) argued that no single
treatment could be recommended or discouraged due to inconsistent stand responses.
Protection of advance regeneration was emphasized to encourage NWC in future stands
(Thornton 1957b). One study suggested that gross growth of NWC is unresponsive to
stand density (Foltz and Johnston 1968); results indicate that NWC stands can be thinned
repeatedly to a basal area 21 m2
/ha without sacrificing gross growth. Partial cutting,
group selection, and diameter-limit cutting have been discouraged in deeryards because
they tend to reduce available browse and cover (Verme 1965). Indeed, the Management
Handbook for Northern White Cedar in the North Central States stresses the need for
managers to consider both timber and wildlife implications in any silvicultural treatment
that is prescribed (Johnston 1977).

25. 11
In the Lake States, NWC management is heavily impacted by a lack of
consistency in NWC regeneration and responses to silviculture, to some extent a result of
local deer populations. Effective implementation of treatments for NWC regeneration and
recruitment are limited by high browse pressure and slow seedling growth rates, as well
as inconsistencies in the literature about NWC’s successional niche and shade-tolerance
(e.g. Curtis 1946). Stand conversion is common because NWC is a poor understory
competitor in comparison to species such as balsam fir. It is seemingly difficult to predict
responses of mature NWC stands to silvicultural treatments; at times they respond well to
release (e.g. Roe 1947), at times they show no difference among treatment intensities
(e.g. Foltz and Johnston 1968). These concerns make management of deeryards and other
NWC stands problematic.
1.3.4 The Northeast
One of the oldest NWC studies on record (Fernald 1919) established an important
link between forest vegetation and soils and underlying parent material in the Northeast.
Fernald (1919) found that while NWC can grow on acidic soils, the best growth and form
were observed on calcareous sites. Furthermore, research comparing bog NWC trees with
limestone outcrop NWC in New York suggested that many wood properties have greater
variance within site than among sites (Harlow 1927). Specific gravity and crushing
strengths were higher in a limestone than bog community, as was growth increment (3.10
mm/yr versus 1.55 mm/yr); though growth differences were partially attributed to
“openness” of the limestone community.
Much of the knowledge about NWC in the Northeast comes from studies in
Maine during the 1940s. NWC was noted to have a spiral grain pattern across all sites,

26. 12
though upland NWC commonly had higher volume growth and better stem form (Curtis
1944, 1946). NWC was thought to be a pioneer species on abandoned pasturelands that
was capable of growing 4.5 m3
/ha/yr in stands exceeding 73 m2
of basal area per hectare.
Fomes roseus and Polyporous schweinitzii were cited as common decay fungi in older
NWC. Curtis (1946) expressed a concern for future NWC stands in his observations of
abundant seedlings but few saplings; a similar phenomenon was noted in western Nova
Scotia (Ringius 1979). Though conventional wisdom suggests that stem quality and
growth increases as site class improves, central decay incidence and proportion of breast
height area decayed were found to increase as soil drainage improved (Hofmeyer et al.
2006). Hofmeyer et al. (2006) also reported that mature NWC basal area growth was
unaffected by canopy position or drainage class and had slightly lower growth than
associated species in the region.
NWC in the Northeast is considered a slow-growing species, attaining maximum
heights of 24 m in 125 years on the best lowland sites (Hannah 2004). Site index was
commonly 8 to 16 m at 50 years in Vermont; the wettest sites on the lower end of that
scale. Volume increment data suggest that one could expect 0.08 m3
per tree on poor sites
and up to 0.4 m3
per tree on better sites in 75 years. In Vermont, Hannah (2004)
speculated that NWC would likely be replaced in the future by more competitive species
on better sites.
1.4 ECOTYPIC VARIATION
Ecotypic variation is defined as genetic variation within populations dispersed
across different environments. The hypothesis that NWC exhibits ecotypic variation was
proposed by Potzger (1941) and has been tested in a number of regions. This is important

27. 13
to management, because guidelines developed for a given site or region may not apply to
others if NWC trees are inherently (genetically) different. NWC is widely recognized to
have a bimodal site distribution, occurring commonly on xeric and hydric sites (e.g.
Potzger 1941, Musselman et al. 1975, Collier and Boyer 1989). Several researchers have
conducted comparative investigations to determine the likelihood that xeric and hydric
communities are part of separate populations.
Seedlings were collected from upland and lowland communities in Wisconsin and
transplanted to the University of Wisconsin Arboretum to observe differences in seedling
morphology (Habeck 1958). NWC seedlings from upland sites were found to have higher
survival rates and more plastic root development than lowland NWC seedlings, leading
Habeck (1958) to conclude that ecotypic variation does exist. A later study in Wisconsin
supported this finding, suggesting that root structures changed from tap roots with few
laterals to no tap root but many laterals as soil moisture levels increased (Musselman et
al. 1975). Another study reports on seedlings (2-0 stock) from 32 widely distributed sites
that were planted as windbreaks in Illinois (Jokela and Cyr 1977). After 12 years, no
survival differences in growth rate or susceptibility to winter foliage damage were
detected among provenances; however, no geographic pattern was observed. The authors
speculated this could be a reflection of localized lowland and upland ecotypes in the seed
stock.
The conclusion that NWC exhibits ecotypic variation has not gone unchallenged,
however. Tree architecture was found to vary little among xeric and hydric sites, once
age was corrected for (Briand et al. 1991). Seed morphology followed a similar pattern:
within-site variability exceeded differences among sites (Briand et al. 1992). Cell water

28. 14
potential, which has been shown to correspond with habitat moisture conditions in many
species, provided another means to test NWC for ecotypic variation (Collier and Boyer
1989). Using hydrated foliage samples in a Pressure Bomb, water potential was found to
be more negative in a xeric moisture regime, independent of parent stock. Investigating
NWC drought response mechanisms, no significant differences were detected in
transpiration rates and osmotic adjustment between well-watered and stress-conditioned
individuals in subsequent moisture stress relief (Edwards and Dixon 1995). Cliff and
swamp populations of NWC were tested for ecotypic differences in growth and tree
physiology (Matthes-Sears and Larson 1991); no discernable patterns of productivity,
nutrient levels, shading, or light saturation were observed among sites. Allozyme patterns
among cliff and swamp sites have also been investigated (Matthes-Sears et al. 1991);
only 1.9 percent of genetic variability distinguished stands within habitat types, and only
1.2 percent detected differences between swamps and cliffs, suggesting that all trees
studied were members of a homogeneous population.
In each of the cases in the preceding paragraph, the authors concluded that no
inherent differences in NWC existed among sites. Many authors that reject ecotypic
variation suggest that NWC is a highly variable species and that variations are site-, not
habitat-, specific. Some authors suggest that this variability may allow NWC to establish
and persist across a wide range of environmental conditions (Briand et al. 1992).
1.5 SUMMARY
Though the depth and focus of NWC research has varied regionally, a number of
generalizations can be made. NWC is often described as a small stature tree rarely
exceeding 25 m in height; it is slow-growing and decay is problematic in larger and/or

29. 15
older individuals (except on cliff sites). Growth and stem form have been reported to be
better on upland sites than on lowland sites. Identifying NWC’s ecological and
successional niche has proven challenging; it is reported both as a pioneer that is replaced
by competing species, and as a shade-tolerant conifer in multi-aged stands. NWC is
generally associated with higher pH and calcareous soils.
Seedling regeneration can be abundant after disturbance; however, seedling
densities often drop quickly due to excessive browsing, desiccation, and slow growth.
Because of this, sapling recruitment is often low, leading to stands with an overstory of
NWC and a competing species in the understory (e.g. balsam fir). This is of particular
concern in areas with harsh winters and high deer populations.
Silviculture research is limited and cutting trials have yielded inconsistent results
regarding response to release. Though some researchers suggest focusing management in
upland stands where growth and stem form are believed to be better, this is not supported
by recent studies showing more decay on upland sites. Though NWC has a bimodal site
distribution, it is not necessarily a species with two subpopulations. Much of the literature
suggests NWC is a highly variable species that can adapt to a wide range of
environmental stresses. NWC will continue to prove challenging to managers and
researchers, but the limits of current knowledge mean that there is great promise for
future discoveries.

30. 16
Chapter 2:
INFLUENCE OF SOIL SITE CLASS ON GROWTH AND DECAY OF
NORTHERN WHITE-CEDAR
2.1 ABSTRACT
Basal area growth of outwardly sound northern white-cedar (Thuja occidentalis
L.) was compared to that of balsam fir [Abies balsamea (L.) Mill.] and red spruce (Picea
rubens Sarg.) on 60 sites throughout northern Maine across site and light exposure class
gradients. Once adjusted for sapwood area, northern white-cedar basal area growth was
not strongly affected by site or light exposure class; growth was similar to that of red
spruce but generally lower than that of balsam fir. Site index did not differ appreciably
among soil drainage classes for red spruce and northern white-cedar, though small
sample size on upland site classes limited analysis. Incidence of central decay was higher
in northern white-cedar than balsam fir, which was higher than red spruce. Incidence of
decay in outwardly sound northern white-cedar and balsam fir was highest on well-
drained mineral soils. Mean proportion of breast height area decayed increased in
outwardly sound northern white-cedar as drainage improved from poorly to well-drained
soils. These data suggest that northern white-cedar on lowland organic and poorly
drained mineral soils in Maine have superior stem soundness, similar basal area growth,
and similar site index relative to upland communities.

31. 17
2.2 INTRODUCTION
Northern white-cedar (Thuja occidentalis L.) is a common tree species in mixed
transitional forests of southeastern Canada and northeastern United States (Johnston
1990). In Maine, it is commonly found in association with balsam fir [Abies balsamea
(L.) Mill.], red spruce (Picea rubens Sarg.), black spruce [Picea mariana (Mill.) B.S.P.],
and eastern larch [Larix laricina (Du Roi) K. Koch]. Northern white-cedar is the third
most abundant conifer in Maine, after balsam fir and red spruce; these species account for
6.3, 21.7, and 12.7 million cubic meters, respectively (McWilliams et al. 2005). Though
northern white-cedar is prevalent on the landscape, it has historically been under
represented in ecology and silviculture research throughout its native range, particularly
in the northeastern United States (Hofmeyer et al. 2007).
Red spruce and balsam fir are commonly associated species in the spruce-fir
forests of Maine. These species occupy similar sites, are both shallow-rooted (though fir
slightly deeper than spruce), very shade tolerant, and can be considered climax species in
that they reproduce under their own shade (Zon 1914, Murphy 1917). The most common
and widespread occurrence of the spruce-fir type occurs on “spruce flats:” relatively
shallow soils extending from swamp sites to lower slopes (Westveld 1931). These soils
are generally moist with perched water tables during the active growing season. Above-
ground growth of spruce and fir is related to site class; growth increases as soil drainage
improves from poorly to well-drained (Williams et al. 1990, Briggs 1994). Little research
has been conducted regarding growth on organic soils because red spruce-balsam fir
forests have less importance on these soil types.

32. 18
Northern white-cedar can occur in nearly pure stands on both upland (e.g.
abandoned pastures) and lowland (e.g. poorly drained mineral and organic soils) sites in
Maine (Curtis 1944, 1946). Northern white-cedar also occurs in mixtures with red spruce
and balsam fir in transitional stands with improved drainage, becoming more widely
scattered as drainage improves to moderately well-drained in mixedwood stands with an
important component of mesic northern hardwoods.
Though there is a wealth of anecdotal evidence suggesting that growth rate of
northern white-cedar is superior on upland soils and abandoned pasture lands (Curtis
1946, Caulkins 1967, Johnston 1990), this claim has not been rigorously tested. In a study
of northern white-cedar on wet sites in Vermont, Hannah (2004) found differences in
height and volume growth between the poorest and best lowland sites. Total yield of
northern white-cedar stands in the Lake States increased as site index increased
(Gevorkiantz and Duerr 1939), suggesting that volume increment increases as site
conditions improve. However, there is a lack of fundamental research on soil-site
relationships of northern white-cedar throughout its native range, and a lack of growth
comparisons to associated species across the soils continuum.
Anecdotal evidence also suggests that in addition to superior growth on upland
sites, northern white-cedar stem quality is better on these sites (Curtis 1946, Johnston
1990). Central decay resulting from heart rot fungi is commonly cited as problematic for
cedar throughout its native range (Harlow 1927, Johnston 1990), with the exception of
stunted northern white-cedar trees growing on limestone cliffs in southern Ontario
(Larson and Kelly 1991). Though decay affects a high proportion of northern white-
cedar, no study has rigorously tested its occurrence by site class.

33. 19
The objectives of this study were to 1) compare basal area growth and decay of
upper-canopy, outwardly sound northern white-cedar to that of balsam fir and red spruce
as a function of site class and canopy position, and 2) compare site index of northern
white-cedar to balsam fir and red spruce by site class.
2.3 METHODS
2.3.1 Site Description
Sixty sites were selected throughout central and northern Maine for this study
(Figure 2.1). Study sites were supplied by ten landowners or land managers, each with
different forest management objectives. Detailed forest history for each site is unknown.
Observation of cut stumps in various stages of decay indicated that most sites had been
partially harvested in the past; exceptions were sites within streamside management
zones and white-tailed deer (Odocoileus virginianus) wintering areas.
Climate in northern Maine is cool and moist. National Oceanic and Atmospheric
Administration long-term climate data indicate a mean annual temperature of 3.95o
C (2.2
- 7.1o
C) and mean annual precipitation of 97.5 cm (90.2 - 105.7 cm) for this region1
.
Glacial retreat ca. 10,000 to 12,000 years ago deposited dense basal till that is the
parent material for most low-lying forests in the region. Soils of this nature with high silt
content commonly have imperfect drainage and poor solum development. Because of the
parent material and cool climate, soils throughout northern Maine are generally young in
terms of weathering, and exhibit physical properties similar to the underlying parent
1
NOAA stations in Allagash, Bangor, Caribou, Clayton Lake, Fort Kent, Jackman, and
Millinocket. Climate data from 1948-2006. Accessed at http://www4.ncdc.noaa.gov/ on
06/29/2007.

34. 20
material. The dominant forest soils in Maine are Spodosols (approximately 65%) and
Inceptisols (29%) (Fernandez 1992). Organic acid leaching from the forest floor leads to
accumulations of iron and aluminum in the subsoil. Upland sites with improved drainage
are typically Orthods (freely drained Spodosols), while areas of poor drainage are
Figure 2.1. Study site locations throughout central and northern Maine.

35. 21
often Aquepts (wet Inceptisols with poorly developed B horizons). Orthods occur on 59%
of the land area in Maine and Aquepts on 24% (Fernandez 1992). In general, these soils
are acidic, high in organic matter, low in base saturation, and nutrient poor. Though
spruce-fir forests in Maine are generally associated with Spodosols and Inceptisols,
northern white-cedar communities tend to be associated with Inceptisols and Histosols.
Poorly drained Hemists occurring in low-lying areas and freely drained Folists resting on
bedrock can support spruce-cedar communities, and occasionally cedar-fir. Northern
white-cedar becomes more sporadic and is frequently replaced by northern hardwoods as
soil drainage improves. Sampled sites in this study focus on red spruce-balsam fir-
northern white-cedar associations on these soil types.
2.3.2 Field Data Collection
Sampling occurred from June 1 through August 30 in 2005 and 2006. Five upper
canopy northern white-cedar, red spruce and/or balsam fir trees were located at each site
and their light exposure class was identified. Light exposure classes were identified
following Bechtold’s (2003) protocol to reduce errors associated with assigning
traditional crown classes in stratified or multi-cohort stands (Nichols et al. 1990). Light
exposure was rated on a 1-5 scale for each tree; class 5 is analogous to a dominant tree
(light on the top and four sides) and class 1 is analogous to an intermediate (light on the
top or one side only). Only trees in the continuous upper canopy with a light exposure > 1
were sampled; no overtopped or outwardly defective trees were sampled. Each sample
tree was double cored to the pith perpendicular to one another at breast height. Cores
were held to the sky to identify the boundary of translucent sapwood and opaque xylem
prior to mounting. Cores were mounted in the field and tree number, light exposure class,

36. 22
sapwood thickness and point of decay (if present) were marked on the core board. Bark
thickness was measured with a bark gauge to the nearest millimeter at each core location
on the bole. Tree diameter was measured at breast height (1.3 m) with a steel diameter
tape to the nearest millimeter. Total height, height of the live crown base, and height of
the lowest live branch were measured with a Haglof Vertex III hypsometer. Live crown
base was defined as the point on the bole with living branches covering at least 50% of
the circumference of the bole.
A centrally located sampling point was taken at each site with a BAF 10 prism to
characterize stand density and species composition. Two soil pits were excavated at each
site to determine site class. Soil pits were located at the edge of the widest portion of the
site and along a topographic gradient where possible (e.g. one uphill and one downhill or
one pit and one mound sampled). Depth to redoxymorphic features, depth to root
restriction, and percentage of coarse fragments were recorded for each pit. Each site was
then placed into the appropriate site class using Briggs’ (1994) classification (Table 2.1).
Sites with soil pits varying by more than one site class were rejected due to excessive soil
variability. Because northern white-cedar is commonly found on organic sites with no
redoxymorphic features present, organic sites were treated as a separate site class. A GPS
point was taken at each site for mapping and relocation purposes.
Mean diameter, height, and breast height age varied by species, though northern
white-cedar had the shortest mean total height, the largest mean outside bark diameter,
and the oldest mean breast height age (Table 2.2). Tree age was determined using only
sound cores. In the event that sound cores did not intersect the pith, transparencies with

37. 23
Table 2.1. Briggs (1994) site class descriptions.
Site
class Drainage class
Depth to
mottling **
Loam cap
thickness
Well drained >24" -
1
Somewhat excessively - >12"
Somewhat excessively - 8-12"
2
Moderately well 16-24" -
3 Somewhat poorly 8-16" -
Poorly drained 4-8" -
4
Excessively drained * - -
5 Very poorly <4" -
* Shallow bedrock (<12”) or coarse sand and gravel
** Depth to seasonal high water table as indicated by low chroma mottles (or grey).
Table 2.2. Mean height, diameter, and breast height age of sample trees.
Total tree height (m)
Min Mean Max SE
Balsam fir (n=171) 10.0 17.4 23.8 0.228
Cedar (n=296) 7.6 15.7 26.6 0.157
Red spruce (n=158) 11.1 18.6 28.5 0.260
Diameter at breast height (cm)
Balsam fir (n=171) 10.7 22.8 37.6 0.419
Cedar (n=296) 13.8 31.7 65.7 0.539
Red spruce (n=158) 12.0 28.2 52.5 0.634
Breast height age (ring count)
Balsam fir (n=134) 23 68.2 151 2.84
Cedar (n=96) 37 129.3 233 3.36
Red spruce (n=137) 42 111.6 200 2.71

38. 24
concentric circles of equal radial width were used to estimate the number of years to the
pith (after Applequist 1958). Sample trees of each species were present at each site and
light exposure class level (Table 2.3).
Table 2.3. Number of sample trees by site and light exposure classes.
Site Class* Balsam fir Northern white-cedar Red spruce Total
2 28 32 10 70
3 30 40 15 85
4 20 50 30 100
5 63 105 58 226
Organic 30 69 45 144
Total 171 296 158 625
LE Class
1 30 64 23 117
2 49 88 30 167
3 48 94 38 180
4 25 36 36 97
5 19 14 31 64
Total 171 296 158 625
* Site class descriptions follow Briggs’ (1994) classification; LE Class – Light exposure
class, after Bechtold (2003).
2.3.3 Data Analysis
Tree cores were dried, sanded, and analyzed with Regent Instruments WinDendro
software (1600 dpi resolution). Cores were analyzed from bark to pith, counting and
dating annual radial increment at the juncture of latewood and earlywood. Sapwood
thickness was measured with digital calipers to the nearest 0.1 millimeter. Four balsam
fir, four northern white-cedar, and two red spruce sample trees were removed from the
analysis due to damaged cores.
Basal area increment was computed for the most recent complete five years of
growth. Both sapwood area and basal area increment were determined at the core level

39. 25
and averaged for tree-level values. A no-intercept least-squares general linear model was
used to describe the relationship of basal area growth as a function of sapwood area.
Analysis of covariance (ANCOVA) was used to test differences in basal area growth
among site classes and light exposures for each species at α=0.05. Sapwood area was
used as a covariate in these analyses because of the allometric relationships between
breast height sapwood area and foliage mass or area in coniferous tree species (e.g. Grier
and Waring 1974, Gilmore and Seymour 1996, Maguire et al. 1998). Independent class
variables were species, site, and light exposure; all independent variables were tested for
interactions. Mean separations were analyzed with Tukey’s HSD test.
Tree decay was investigated with two metrics: the proportion of trees with central
decay, and the proportion of basal area that was centrally decayed. Trees were considered
“decayed” if either of the two cores showed evidence of decay (i.e. cells decomposed)
before the pith. Proportion of basal area decayed was quantified as:
[1] Decayed area (m2
) = 00007854.02
2
2
×⎥
⎦
⎤
⎢
⎣
⎡
×
⎭
⎬
⎫
⎩
⎨
⎧
−⎟
⎠
⎞
⎜
⎝
⎛
CL
DBHib
where DBHib is the diameter inside bark (cm) and CL is the core length (cm). Decayed
area was quantified for each core and averaged over both cores to determine the tree
basal area decayed. The mean value from [1] was divided by the inside bark basal area to
determine the proportion of basal area that was decayed. Analysis of variance (ANOVA)
was used to test for differences in the proportion of basal area decayed by site class and
light exposure class for each species at α=0.05.
Tree cores with no evidence of decay were selected for site index analysis. Site
index trees should be dominant or codominant and free from suppression (Avery and

40. 26
Burkhart 1994). Light exposure class 1 and 2 trees were eliminated from this analysis.
Chronologies were made for each tree that was free from decay. Only individuals with no
signs of early suppression, as evident by reduced radial increment in the chronology,
were retained for this analysis (see Appendix A). Nearly all of the red spruce and balsam
fir trees in this study had spruce budworm (Choristoneura fumiferana) signals in their
ring chronologies. To increase sample size, several trees with weak budworm signals
were retained in the analysis. A five-parameter Weibull function was used to estimate site
index with parameter coefficient estimates previously published by Carmean (1989) for
northern white-cedar and Steinman (1992) for red spruce and balsam fir. ANOVA was
used to test site index differences among site classes for each species at α=0.05. SYSTAT
version 12 was used for all statistical analyses.
2.4 RESULTS
2.4.1 Basal Area Growth
A general linear model of basal area growth as a function of sapwood area had
significant slopes for all species, and accounted for greater than 80% of the growth
variability (Figure 2.2). Red spruce had the highest proportion of sapwood area per unit
basal area (SA:BA) (mean=0.297, SE=0.007), followed by balsam fir (mean=0.252,
SE=0.006) and northern white-cedar (mean=0.178, SE=0.005). SA:BA differed by light
exposure class only in northern white-cedar, with a significant trend of decreasing
SA:BA with increased light exposure (p=0.006, Table 2.4).
Sapwood area was a significant covariate (p<0.001) for tests of basal area growth
differences among site and light exposure classes by species. Site class was not a
significant predictor of basal area growth in any species (Table 2.5). Balsam fir exhibited

41. 27
increased basal area growth with increasing light exposure (p=0.004); however, neither
northern white-cedar nor red spruce showed a growth response to increased light levels
(p=0.881 and p=0.425, respectively). As expected, sapwood area and light exposure class
were highly correlated in each species and may have masked differences in basal area
growth among light exposure classes. Sapwood area did not differ across site classes
within species.
2.4.2 Decay
Mean incidence of decay (across all site classes) was lowest in red spruce (11% of
the sample), followed by balsam fir (34%) and northern white-cedar (80%). Balsam fir
and northern white-cedar incidence of decay was higher on site class 2 than on site class
5 and organic sites; red spruce decay was not different among site classes (Table 2.6).
Site class did not affect area decayed in balsam fir (p=0.092) or red spruce (p=0.273), but
was significant for northern white-cedar (p<0.001) (Figure 2.3). Northern white-cedar
had a trend of increasing area decayed with improved soil drainage. Light exposure class
did not influence decay in any tree species (p>0.20).
2.4.3 Site Index
Surprisingly, northern white-cedar and red spruce site index did not differ among
Briggs’ (1994) site classes (p=0.641 and p=0.358, respectively). Balsam fir site index
was higher in site class 4 (p<0.001); all other site classes were undifferentiated. Balsam
fir was taller at breast height age of 50 years than red spruce and northern white-cedar in
most site classes (Table 2.7). Though site index of red spruce and northern white-cedar
were not significantly different from one another across site classes, red spruce generally

42. 28
10
20
30
40
50
Basalarea(cm2/yr)
10
20
30
40
50
Sapwood area (cm2)
0 200 400 600 800
0
10
20
30
40
50
Balsam Fir
BA growth = 0.084*SA
r2 = 0.80
Northern white-cedar
BA growth = 0.066*SA
r2 = 0.88
Red spruce
BA growth = 0.046*SA
r2 = 0.84
Figure 2.2. Basal area (BA) growth as a function of sapwood area (SA) by species.

43. 29
Table 2.4. Northern white-cedar sapwood area per unit basal area (SA:BA) by light
exposure class.
Light exposure class n Mean SA:BA SE
1 65 0.196 0.007
2 88 0.181 0.006
3 93 0.172 0.006
4 36 0.163 0.009
5 14 0.151 0.015
Table 2.5. Mean basal area growth as a function of site class and light exposure class with
a significant sapwood area covariate.
Balsam fir Northern white-cedar Red spruce
Site Class mean* SE mean SE mean SE
2 8.76 0.90 9.32 0.64 9.96 1.44
3 9.72 0.86 8.24 0.57 9.91 1.18
4 9.91 1.06 8.23 0.51 10.58 0.83
5 9.10 0.60 9.18 0.35 8.68 0.60
Organic 7.36 0.86 9.05 0.43 9.06 0.68
LE Class mean SE mean SE mean SE
1 6.14b
0.91 8.83 0.50 8.35 1.03
2 8.37ab
0.66 8.84 0.39 10.14 0.86
3 9.41ab
0.66 8.67 0.38 9.51 0.74
4 11.34a
0.93 9.42 0.63 9.93 0.78
5 10.51a
1.15 9.21 1.01 8.43 0.86
* Mean growth is in cm2
/year.
Note: means followed by different letters are different at the α=0.05 significance level
(Tukey’s HSD mean separation).

44. 30
Table 2.6. Mean incidence of sample trees decayed by site class (standard errors in
parentheses).
Proportion of sample decayed
Site Class Balsam fir Cedar Red spruce
2 0.57 (0.09)a
0.97 (0.07)a
0.10 (0.10)
3 0.40 (0.08)ab
0.88 (0.07)ab
0.13 (0.08)
4 0.40 (0.10)ab
0.64 (0.06)b
0.13 (0.06)
5 0.19 (0.06)b
0.73 (0.04)b
0.07 (0.04)
Organic 0.23 (0.08)b
0.74 (0.05)b
0.13 (0.05)
Mean 0.34 (0.03)c
0.80 (0.03)d
0.11 (0.02)e
Note: Means of the same species followed by different letters are different at α=0.05
significance level (Tukey’s HSD mean separation). Species’ means were also
significantly different.
Site class
2 3 4 5 org
Proportionofbasalareadecayed
-0.05
0.00
0.05
0.10
0.15
0.20
Cedar (p<0.001)
Balsam fir (p=0.092)
Red spruce (p=0.273)
Figure 2.3. Proportion of sample tree basal area decayed by site class in northern white-
cedar, balsam fir, and red spruce.

45. 31
Table 2.7. Site index (m) at breast height age 50 for balsam fir, northern white-
cedar and red spruce by site class determined from trees without radial growth
suppression in light exposure classes 3 through 5 (standard errors in parentheses).
Site Class Balsam fir Cedar Red spruce p-value
2 14.8 (1.188)ab
12.1 (1.372) 0.196
n 4
no data
3
3 14.6 (0.687)ab
12.3 10.9 (1.023) 0.015
n 7 1 2
4 17.2 (0.956)a
10.3 (0.956) 12.5 (1.022) <0.001
n 8 8 7
5 13.2 (0.334)b
10.5 (0.732) 10.6 (0.437) <0.001
n 24 5 14
Organic 13.6 (0.434)b
9.6 (0.614) 11.2 (0.367) <0.001
n 10 5 14
p-value <0.001 0.091 0.359
Note: Column means followed by differing letters were different at the α=0.05 level
(Tukey's HSD mean separation).

46. 32
had a higher mean site index. Small sample size reduced confidence in site index values
on upland site classes.
2.5 DISCUSSION
2.5.1 Basal Area Growth
Basal area growth is often difficult to interpret because it is heavily influenced by
tree size, crown variables, and stand density. Sapwood area was used in this study as a
surrogate for leaf area, which in turn reflects stand density and associated crown variables
(Grier and Waring 1974, O’Hara 1988). Mean SA:BA results indicated that northern
white-cedar had less sapwood area than red spruce and balsam fir for a given tree
diameter, which is consistent with past reports that cedar has a narrow sapwood radius
per given diameter (Curtis 1946, Behr 1974). Linear regression analysis indicated that
northern white-cedar basal area growth per unit sapwood area is less than balsam fir, but
more than red spruce (see Figure 2.2). Because of the high proportion of sapwood area
and the strong influence of basal area growth per unit sapwood area, balsam fir appears to
have the best basal area growth capabilities of the three species at a given diameter and
canopy position. High correlation of sapwood area to light exposure class suggests
favoring trees in superior canopy positions for all species because basal area increment
increases with sapwood area.
Basal area growth results were somewhat surprising with regard to light exposure
and site class. Because of the correlation between sapwood area and light exposure class,
much of the growth response was captured in the covariate. When sapwood area was
removed from the analysis, all three species showed a strong significant increase in basal

47. 33
area increment at higher light exposures (p<0.01). This is consistent with expectations of
higher growth increment of trees in superior canopy positions (Assman 1970).
On organic soils, balsam fir growth was significantly lower than that of red spruce
and northern white-cedar. This is consistent with Briggs and Lemin (1994), who found
that growth response of balsam fir to precommercial thinning was lowest on poorly and
excessively drained soils. Meng and Seymour (1992) found that balsam fir saplings
expressed significantly higher height and area growth in response to herbicide release
treatment on well-drained site classes relative to poorly drained sites. Their data indicated
that red spruce is relatively unresponsive to this cultural treatment by site class; height
increment and area growth were lower in red spruce than balsam fir on well-drained sites.
We expected basal area growth of northern white-cedar to have a strong positive
correlation with site class, but that relationship was not supported. Godman (1958) and
Caulkins (1967) suggested that cedar growth was higher on upland soils and abandoned
pasture lands than swamp sites in the Lake States. Harlow (1927) reported higher mean
annual radial increment from a cedar community on a limestone outcrop (3.10 mm/yr)
than from a bog community (1.55 mm/yr) in New York, but this difference was partially
attributed to the “openness” of the limestone community. Curtis (1946) reported that
volume growth was better on upland sites than lowland sites in Maine. In the present
study, no differences were detected among upland and lowland communities once crown
variables were accounted for with the sapwood area covariate.
It is important to note that our findings pertain only to basal area growth at breast
height, and may not apply to volume growth. Variability in height and radial increment
make volume predictions based solely on area increment tenuous. Nevertheless, breast-

48. 34
height area increment is a commonly used metric of tree growth; results of the present
study thus contribute meaningfully to our understanding of northern white-cedar
stemwood growth dynamics.
2.5.2 Stem Decay
Species differences in mean breast height area decayed observed in this study are
consistent with reports of high incidence of decay in northern white-cedar (Harlow 1927),
moderate incidence of decay in balsam fir and little decay in spruce (Whitney 1989).
However, the trend in northern white-cedar for higher proportions of sample tree breast
height area to be decayed on upland sites was somewhat unexpected. Several studies
throughout the Northeast and Lake States suggest that stem quality is superior on upland
sites (e.g. Curtis 1944, Godman 1958, Johnston 1990). Though “stem quality” is often
ambiguously defined in cedar literature and might not be specific to internal decay, in
terms of soundness, stem quality on well-drained sites in central and northern Maine is
inferior to stem quality on very poorly drained and organic sites.
Results indicating no decay differences among light exposure classes suggest that
cedar trees occupying all upper canopy positions are equally susceptible to decay fungi.
Because no overtopped trees were sampled in this study, it remains unknown the degree
to which canopy suppression is correlated with northern white-cedar’s decay fungi
vulnerability. Reduced photosynthate and nutrient allocation to defensive compounds has
been suggested to cause an increase in vulnerability to decay fungi (Waring 1987).
There was a significant site effect on decay incidence and proportion of sample
tree basal area decayed. Several species of decay fungi have been observed to affect
northern white-cedar including Armillaria mellea, Phaelus schweinitzii, and

49. 35
Heterobasidion annosum (Hepting 1971, Johnston 1990). P. schweintzii and A. mellea
have been reported as common decay fungi infecting balsam fir in the Northeast. Basham
et al. (1953) reported that butt rot was more common in balsam fir on upland
“mixedwood slopes” than on “softwood flats.” This relationship was attributed to fungus
site preferences and the absence of many fungal species on the lower site classes. A
similar site-decay relationship was reported earlier by Zon (1914); greater balsam fir
decay on upland sites was attributed to frequent anaerobic conditions on lowland sites
that limited fungal infection. Similar site influences on decay in balsam fir and northern
white-cedar in this study support the notion that decay fungi are more prevalent and
vigorous on upland sites.
Wood strength might play an important role in decay fungi entry and proliferation
within the bole. Though cellulose is essentially the same in all trees, differences in
hemicellulose and lignins do occur and impact tree responses to decay fungi (Manion
1991). Northern white-cedar is the lightest and weakest commercial wood in the United
States; it has a density of 0.315 g/cc and a modulus of rupture of 4.56 kg/mm2
compared
to balsam fir (0.414 g/cc, 5.42 kg/mm2
) and red spruce (0.414 g/cc, 7.15 kg/ mm2
) (Seely
2007). Some evidence suggests that increased wood density provides some protection
against pathogens (Loehle 1996). One might hypothesize from these metrics that decay
would be more prevalent in northern white-cedar than red spruce and balsam fir. Whitney
(1989) found that decay fungi were more common in balsam fir than black spruce
(density of 0.428 g/cc, modulus of rupture of 7.24 kg/ mm2
). Whitney (1989) observed
high incidences of root decay in young balsam fir stands, perhaps as a result of the
inability of balsam firs’ weak wood to withstand environmental stressors (e.g. wind and

50. 36
snow). Decay results from this study suggest that decay and wood strength may be
correlated in some species.
All sites selected for sampling were provided by members of forest industry with
active harvesting operations. On many upland sites, forest operations preferentially
removed balsam fir, red spruce and high value hardwoods. Northern white-cedar is often
retained on these sites to meet minimum stocking levels post harvesting. Residual stand
damage is common after partial harvesting (Ostrofsky et al. 1986, Ostrofsky and Dirkman
1991). Northern white-cedar may be particularly susceptible to crown and root damage
during harvesting because branches and roots are weak. If the branch collar remains
intact during crown damage events, decay generally cannot spread throughout the bole
(Shigo 1984). If a branch is stripped away due to wind, snow load, or harvesting and
damages the branch collar, decay could enter the stem.
Worrall et al. (2005) noted that balsam fir has a high incidence of root damage
due to chronic wind stress in fir waves in New Hampshire. Many of the stems in that
study were observed to be infected by Armillaria and other root rot fungi. Northern
white-cedar stands commonly occur on shallow or poorly drained sites where individuals
occupy soil mounds and decayed woody material that were once germination sites. These
stems are frequently “pistol butted” due to a disturbance of the root system (e.g.
harvesting operations, snow loads, and wind stress). Zon (1914) reported decay fungi
entry through root damage from sharp rocks, strong winds, and logging operations in
balsam fir. If decay fungi enter disturbed root systems in northern white-cedar as they do
in fir, it seems that root breakage could be an entry mechanism.

51. 37
Core data suggest differences in ages among the sampled species. These data
indicate that, on average, the sampled mixed-species spruce-fir-cedar stands were multi-
aged. Red spruce is an inherently long-lived species that can attain ages exceeding 300
years in virgin forests (Cary 1894); though the mean age of red spruce in this study was
far younger than its maximum, it was older than the mean balsam fir age. Balsam fir is a
species with a pathological rotation set by heart rot and spruce budworm to 40 to 70 years
(Seymour 1992). However, age data from sound cores may be a biased estimation of
mean stem age in these species mixtures. Tian and Ostrofsky (2007) found that there was
a significant increase in balsam fir decay in older stems; this relationship was not
significant in red spruce.
Some disagreement exists among researchers with regard to age characteristics of
northern white-cedar. Larson (2001) reported a typical lifespan of 80 years for non cliff-
dwelling cedar. Johnston (1990) reported a maximum lifespan of 400 years and suggested
that it frequently lives longer than associated species on lowland sites. Age data from the
present study suggest that northern white-cedar on the Maine landscape is generally older
than its associates in mixed-species stands. Mean age of the sound trees clearly exceeded
80 years and the age structure on these sites reflects historical disturbance patterns
throughout the region.
2.5.3 Site Index
Several red spruce and northern-white cedar trees that were selected for site index
analysis exceeded 100 years of age. Given the asymptotic nature of height growth with
stem age, inclusion of these samples might have increased within-site variability, though
residual analysis did not indicate any outliers. Trees were selected for site index analysis

52. 38
post-hoc after the field data collection period. Sample trees could not be assessed for
minor crown damage or crown irregularities that may have impacted total tree height.
Though trees with significant radial increment suppression were removed, it is difficult to
determine the degree of height suppression from minor radial suppression events.
High variability of height growth within site class has been noted for red spruce,
often masking clear differences among site classes. Seymour and Fajvan (2001) reported
mean site index (at stump height age 50) for previously suppressed red spruce trees to
range from 12.4 to 11.3 m on good to poor soil classes, but suggested that high variability
may have been associated with the broad classes they employed. Williams et al. (1991)
reported mean site index (breast height age) for red spruce within a single catena as 17.4
to 15.1 m at 50 years on well-drained to poorly drained sites, while mean balsam fir site
index ranged from 18.0 to 14.6 m. Restricting sampling in that study to a single catena
may have reduced inherent soil variability and associated height growth variation. Red
spruce mean site index values in this study ranged from 12.5 to 10.6 m; within-site
variability was high. Many useable site index trees in this study were on Briggs’ (1994)
site class 4, 5, and organic soils, which makes comparisons with related studies
problematic. Williams et al. (1991) did not sample trees on very poorly drained mineral
or organic soils. Height variability among site classes in the present study could result
from using drainage class to predict growth instead of chemical soil properties, though
Steinman (1992) found that chemical properties accounted for little height variability
after adjusting for physical soil properties in even-aged spruce-fir stands in Maine.
Though this study captured few differences in site index by site class within
species, differences among species were detected. These data suggest that cedar will

53. 39
generally have lower site index than balsam fir along the site class continuum. Williams
et al. (1991) also found that balsam fir has a higher site index than red spruce on better
drained soils, though they reported an opposite relationship on imperfectly drained soils.
2.6 CONCLUSIONS
Conventional wisdom guiding northern white-cedar management throughout the
Northeast had little quantitative support from this study. Field foresters often attribute
higher growth rates and stem quality to upland cedar communities. Results suggest that
cedar basal area and height growth were not affected by site class, though incidence of
decay and proportion of basal area decayed were highest on the well-drained sites.
Though mechanisms for decay entry and physiology of decay responses in
northern white-cedar are largely unknown, efforts to reduce residual stand damage during
harvesting operations are encouraged due to northern white-cedar’s weak and brittle
wood properties. This should help to reduce crown and root disruptions in residual trees.
Avoiding residual stand damage is important because northern white-cedar tends to be
older than associated species in mixed stands and is commonly retained through several
partial harvesting operations.
Northern white-cedar has historically been considered a slow-growing species
relative to many of its competitors. Our study suggests that mature upper canopy northern
white-cedar growth is comparable to red spruce and slightly less than balsam fir,
regardless of site class or light exposure class. This may be particularly important
because northern white-cedar commonly occupies a canopy position inferior to red
spruce and balsam fir due to its small stature. Site index results from imperfectly drained

54. 40
soils indicate that on average balsam fir is taller than red spruce, which is taller than
northern white-cedar.

55. 41
Chapter 3:
LEAF AREA PREDICTION MODELS AND STEMWOOD GROWTH
EFFICIENCY OF THUJA OCCIDENTALIS L. IN MAINE
3.1 ABSTRACT
Stem and crown data from 25 destructively sampled northern white-cedar (Thuja
occidentalis L.) trees were analyzed to estimate projected leaf area (PLA), crown foliage
mass (CFM), annual stemwood volume increment (VINC), and growth efficiency (GE,
volume increment per unit leaf area or foliage mass) over site and light exposure
gradients. PLA was best predicted with a nonlinear model using breast height basal area
and a modified live crown ratio. Allometric PLA, CFM and volume equations were
applied to 256 cored northern white-cedar trees from 60 sites throughout northern Maine
to examine growth efficiency trends. Stem volume was estimated with Honer’s (1967)
equation form with refitted parameters. The relationship of VINC to PLA and CFM was
best predicted by a nearly linear two-parameter power function. Analysis of variance
detected no differences in GE among soil site, light exposure, or decay classes; age was
also not a significant covariate. Northern white-cedar is a stress-tolerant tree species that
is highly adaptable to its immediate surroundings. Large-crowned individuals can be
retained in partial harvests without sacrificing volume increment regardless of soil
drainage, light exposure, or tree age.

56. 42
3.2 INTRODUCTION
Ecophysiological measures of forest stand productivity have been quantified for
several tree species in the forests of the Northeast as stemwood volume increment
(VINC) per unit leaf area (Gilmore and Seymour 1996, Maguire et al. 1998, Seymour and
Kenefic 2002, Innes et al. 2005). The concept of growth efficiency (GE) was proposed by
Waring et al. (1980) as biomass of stemwood growth per unit foliage, and has been used
to describe productivity differences by crown structure or canopy position (e.g. Smith and
Long 1989, Long and Smith 1990, Roberts and Long 1992, Gilmore and Seymour 1996),
stand structure (O’Hara 1988, Maguire et al. 1998, Mainwaring and Maguire 2004), and
site resources (Binkley and Reid 1984, Vose and Allen 1988, Velazquez-Martinez et al.
1992, Jokela and Martin 2004).
Northern white-cedar (Thuja occidentalis L.) is the third most abundant tree
species in Maine behind balsam fir (Abies balsamea (L.) Mill) and red spruce (Picea
rubens Sarg.) (McWilliams et al. 2005). It is often difficult to predict responses of
northern white-cedar to silvicultural treatments in part due to disagreement concerning its
ecological niche and shade tolerance, which has been suggested to range from less than
eastern white pine (Pinus strobus L.) to only slightly less than balsam fir (Curtis 1946). It
has been observed as a colonizing species on dune sites in Michigan that gets replaced by
tolerant hardwoods (Scott and Murphy 1987), a stable component of mixed-species
lowland sites (Kangas 1989), and a long-lived monoculture species in uneven-aged cliff
forests (e.g. Larson and Kelly 1991). One potential method to determining the shade
tolerance and site occupancy of mature trees is to observe differences in GE with respect
to projected leaf area (PLA) or canopy position. Two conceptual models proposed by

57. 43
Roberts et al. (1993) suggest intolerant trees would have a monotonic decreasing GE with
increasing PLA, while shade-tolerant trees would likely have a maximum GE at low to
mid PLA values.
The influence of soil-site relationships on aboveground stemwood growth
efficiency is not fully understood. It has been proposed that higher volume production on
better sites could be from increased foliar efficiency, though this has rarely been
demonstrated except for short periods of time after fertilization trials (Brix and Mitchell
1983, Binkley and Reid 1984, Vose and Allen 1988, Jokela and Martin 2000). Matthes-
Sears et al. (1995) found that nutrient additions increased biomass of northern white-
cedar seedlings, but this was attributed to increased leaf area rather than increased foliar
efficiency.
Northern white-cedar is known throughout its native range to be particularly
susceptible to heart rot fungi (Johnston 1990). With the exception of cliff sites in
southern Ontario, decay is problematic on both limestone outcrop sites and bog
communities (Harlow 1927); however, much of the past research associated better stem
quality with upland sites (Curtis 1946, Johnston 1990). GE has been used as an indicator
of tree vigor and susceptibility to pathogens (Waring 1987, Rosso and Hansen 1998).
Evidence from Hofmeyer (2008, Ch. 2) indicates that internal decay is most prevalent in
northern white-cedar on upland sites in Maine; the degree to which internal decay is
correlated with reduced GE in this species is unknown.
Northern white-cedar ecology and silviculture have been largely ignored in the
literature. The objectives of this study were to develop 1) northern white-cedar tree-level

58. 44
leaf area prediction models, and 2) test the hypothesis that tree GE varies by site class
and canopy position.
3.3 METHODS
3.3.1 Field Methods
Twenty-five northern white-cedar trees were selected from the 296 cored
individuals described in Hofmeyer (2008, Ch.2). Sample trees were selected from a total
of 13 sites distributed throughout northern Maine (Figure 3.1). Sampling was stratified
over Briggs’ (1994) site classes in proportion to the total number of sites sampled. Two
soil pits were excavated at each site to determine site class. Soil pits were located at the
edge of the widest portion of the site and along a topographic gradient where possible
(e.g. one uphill and one downhill or one pit and one mound sampled). Depth to
redoxymorphic features, depth to root restriction, and percentage of coarse fragments
were recorded for each pit. Each site was then placed into the appropriate site class using
Briggs’ (1994) classification (Table 2.1). Sites with soil pits varying by more than one
site class were rejected due to excessive soil variability. Because northern white-cedar is
commonly found on organic sites with no redoxymorphic features present, organic sites
were treated as a separate site class.
Attempts were made to stratify the sample over light exposure classes (LEC) 1
through 5; however, only a single observation of LEC 5 was obtained (Table 3.1). Light
exposure classes were identified following Bechtold’s (2003) protocol to reduce errors
associated with assigning traditional crown classes in stratified or multi-cohort stands
(Nichols et al. 1990). Light exposure was rated on a 1-5 scale for each tree; class 5 is
analogous to a dominant tree (light on the top and four sides) and class 1 is analogous to

59. 45
an intermediate (light on the top or one side only). Only trees in the continuous upper
canopy with a light exposure > 1 were sampled; no overtopped or outwardly defective
trees were sampled. Ideal sample trees were selected when possible (i.e. free from heart
rot, no forking of the bole, no obvious crown damage). Due to high incidence of decay
and poor crown form, several sample trees had some defect. Trees were destructively
sampled between 18 July and 17 August 2006.
Crown projection was measured along six crown radii before felling. Stump
height (0.3 m) and breast height (1.3 m) were marked on the bole and diameter was
measured to the nearest 1 mm. Trees adjacent to the sample tree were felled if they might
hang or cause crown damage to the sample tree. Slash from felled trees was used to
soften the fall and minimize crown damage of the sample tree during felling.
Sample trees were felled at a point along the bole between 0.4 and 1.0 m. A
fiberglass measuring tape was fixed to the stump and run to the leader along the bole such
that the tape was correctly aligned with the 1.3 m mark. Distance to lowest live vigorous
branch, base of the live crown (point on the bole with living branches covering at least
50% of the circumference of the bole), and total height were measured. The tree crown
was divided into three unequal sections; the top half, the mid quartile and the lower
quartile (after Gilmore and Seymour 1997). A branch from each section was chosen at
random based on distance within the crown section. If the randomly chosen branch was
damaged during felling, attempts to reconstruct the branch were made. If reconstruction
efforts were unsatisfactory, the branch was replaced by a second random sample.
Distance along the bole and diameter just beyond the branch collar were recorded for
each sample branch. Five sub-dominant foliar sprays were selected from each sample

60. 46
Figure 3.1. Site locations of the 25 destructively sampled northern white-cedar trees.

61. 47
Table 3.1. Attributes of 25 destructively sampled northern white-cedar trees.
Tree DBH (cm) HT (m) CL (m) Site LEC Age
6 15.9 13.57 6.36 Organic 1 78
9 41.7 19.25 16.12 Organic 3 >148
61 33.0 19.24 10.15 3 3 103
63 23.7 16.76 10.93 3 1 130
71 29.0 17.02 7.43 Organic 1 132
73 36.7 18.29 16.29 Organic 3 186
151 46.3 19.74 16.05 2 2 >105
161 24.7 14.26 8.76 Organic 2 108
165 29.3 16.49 12.54 Organic 4 105
176 34.3 16.20 10.72 4 1 228
180 41.2 15.92 9.19 4 4 203
186 21.9 13.51 9.43 5 1 94
187 14.1 10.53 6.24 5 1 75
188 31.4 14.87 11.53 5 3 142
211 21.0 13.13 11.34 5 1 93
215 26.0 16.23 8.15 5 3 132
407 39.9 18.31 14.92 2 4 >128
408 31.8 15.86 8.18 2 5 176
576 19.1 10.58 5.54 4 3 112
578 24.6 11.00 7.55 4 3 107
589 42.7 19.96 14.41 5 4 202
590 23.0 14.90 9.52 5 1 136
598 24.0 13.62 9.32 3 1 156
600 29.5 12.34 10.4 3 3 101
609 43.5 17.43 10.36 2 3 >118
Table abbreviations: DBH- diameter at breast height, HT- total tree height, CL- length of
the live crown, Site- Briggs (1994) site class, LEC- light exposure class (Bechtold 2003),
Age- number of annual rings at breast height. Sample trees with “>” preceding the ring
count had central decay in excess of 2.5 cm in radial length and could not be confidently
estimated.

62. 48
branch. Sprays were selected from throughout the total range of foliage locations and
morphologies on each sample branch because aging northern white-cedar foliage is
difficult due to the lack of bud scars (Reiners 1974). Sprays were placed into plastic
freezer bags and stored in a cooler on ice. Remaining sample branches were sectioned
and placed into paper bags for air drying. All remaining live branches were measured for
their distance along the bole and basal diameter above the branch collar to the nearest 0.1
mm with digital calipers. Sample trees were then limbed and each successive 1-m interval
was marked on the bole beyond the 1.3 m mark that was previously made.
Cross-sectional discs (approximately 2 cm thick) were removed at each 1 m
interval marked on the bole. An additional disc was taken just beneath the lowest live
branch if this location did not coincide with the standard sampling interval. Bark
thickness of each disc was measured in the field along two radii. Sapwood thickness was
measured in the field along six radii on the lowest live branch and breast height discs.
The interface of the opaque heartwood and the translucent sapwood was readily observed
by holding the disc toward the sky.
Foliage samples stored in coolers were placed into a -15o
C freezer upon return
from the field. Branch sections stored in paper bags were placed in a drying room for five
weeks and dried to a constant mass.
3.3.2 Laboratory Procedure
Dried branch samples were sorted into cones, photosynthetic and non-
photosynthetic tissues. Because northern white-cedar does not have discrete foliage and
woody tissue junctions (Briand et al. 1992), separation of these parts involved some
judgment. Two people sorted all of the foliage samples after a separation protocol was

63. 49
developed, thus reducing sampling bias. To be considered photosynthetic, greater than
75% of the area must have been green. This excluded main axes on some higher order
branches that were predominantly brown woody tissue with few green foliage scales.
Drying did not appear to alter branch tissue color. Dry mass of cones, foliage and woody
tissue was determined to the nearest 0.01 gram on a digital balance.
Frozen foliage samples were scanned using Regent WinSeedle software within 15
minutes of removal from ice. One-sided projected leaf area (mm2
) was determined with a
flatbed scanner on 800 dpi resolution in Regent Instruments WinSeedle software. This
created a black and white image to calculate surface area of the scanner bed covered with
foliage. After scanning the samples, they were placed into a paper envelope and dried at
60o
C for 72 hours. Mass of dried foliage was determined on a digital balance to the
nearest 0.0001 mg within 30 seconds of removal from the oven. Specific leaf area (SLA)
was determined for each foliage sample as the fresh foliage area per unit dry foliage mass
(cm2
/g). Dried SLA foliage mass was added to the dried branch foliage mass to determine
the branch foliage mass (BFM, g) for each branch sample. Branch SLA was multiplied by
BFM to determine branch leaf area (BLA) (Table 3.2).
Cross-sectional discs were surfaced with a drum sander to achieve constant
thickness and were sanded repeatedly with progressively finer grit sandpapers. Two radii
were marked with a pencil on each disc and analyzed using Regent WinDendro software
and a flatbed scanner at 1200 dpi resolution. Six radii were marked and analyzed on each
of the breast height and lowest live branch discs. Tree number, path number, disc height,
and annual radial increment were recorded for each disc. Data files from Regent
WinDendro software were imported into Regent WinStem software for incremental

64. 50
Table 3.2. Mean specific leaf area, branch foliage mass, and branch leaf area for the
lower, mid, and top crown sections of 25 northern white-cedar sample trees.
Crown section Min Mean Max SE
Specific Leaf Area (cm2
/g)
Lower 41.62 61.68 79.44 1.86
Mid 41.76 55.59 68.53 1.44
Top 31.81 46.02 68.85 1.94
Branch Foliage Mass (g)
Lower 29.45 230.28 657.02 32.78
Mid 10.80 170.30 559.61 31.90
Top 1.55 100.77 369.17 17.61
Branch Leaf Area (cm2
)
Lower 2041.59 14257.02 42765.61 2187.72
Mid 538.37 9313.18 29656.56 1698.85
Top 55.79 4549.61 17083.46 770.93
diameter, incremental height, and volume analysis. WinStem calculates stem volume
(dm3
) with an additive cone volume function:
[1] cone volume = (R1
2
+R1*R2+R2
2
)*H*π/3
[1a] tree volume = Σ(cone volume/1000)
where R1 is the radius at one end of the cone (mm), R2 is the radius at the other end of the
cone (mm), and H is the height of the cross-sectional disc. Function [1] is calculated for
each 1-m section, and summed for the tree volume.
3.3.3 Statistical analysis
Analysis of variance (ANOVA) was used to test for SLA differences among
crown sections (lower, mid, top), light exposure classes, and site classes (α=0.05). Three

65. 51
regression models were investigated to predict branch leaf area (BLA) and branch foliage
mass (BFM):
[2] LN(y) = b0 + b1D
[2a] y = (b1Db2
) * (RDb4-1
) * (EXP-(b3RDb4
))
[2b] LN(y) = b0 + b1LN(D) + b2LN(RD) + b3RD
where the dependent variable y is either projected leaf area (BLA, cm2
) or branch foliage
mass (BFM, g), D is the branch basal diameter (mm), and RD is the relative distance of
the branch into the crown. RD ranges from 0 to 1; 0 is the crown leader, 1 is the lowest
live branch. The summation of BLA for all branches of each tree is the tree-level
projected leaf area (PLA, m2
). The summation of BFM for all branches is the tree-level
total crown foliage mass (CFM, kg). Models were compared by their generalized r2
(after
Kvalseth 1985), Furnival’s (1961) index of fit (FI), and residual analysis.
Model [2a], a modified Weibull function, has been used to predict leaf area in
white pine (Pinus strobus L.) (R.S. Seymour, unpublished data), balsam fir and red
spruce (Meyer 2005), and Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) (Maguire
and Bennett 1996). The latter study used a dependent variable of BLA, as opposed to the
square root of BLA used in this study. Model [2b] has been used to determine red spruce
PLA in uneven-aged stands in Maine (Maguire et al. 1998).
A series of linear and non-linear, weighted and unweighted, transformed and
untransformed models were fit to the 25 tree-level PLA and CFM values (Table 3.3).
Many of these models were screened in Meyer (2005) for thinned and unthinned stands
of red spruce and balsam fir. Three main effect weights were screened for each
untransformed model (x-1
, x-2
, x-3
). Model weights were screened to reduce the influence

66. 52
Table 3.3. Projected leaf area (PLA) and crown foliage mass (CFM) models fit to 25
northern white-cedar sample trees (only PLA shown as dependent variable).
Model Model form Source *
AIB1 PLA=b0+b1AIBBH 13
AIB2 PLA=b0+b1AIBLLB 13
AIB3 PLA=b1AIBBHb2
8, 10, 13
AIB4 PLA=b1AIBLLBb2
1
AIB5 PLA=b1AIBBH*mLCR 13
AIB6 PLA=b1AIBBHb2
*mLCR 8, 10, 13
AIB7 PLA-b1AIBBH*mLCRb2
13
AIB8 PLA=b1AIBBHb2
*mLCRb3
13
AIB9 PLA=b1AIBBHb2
*CL 13
AIB10 PLA=b1AIBBH*LCRcb -
AIB11 PLA=b1AIBBH*LCRllb 13
AIB12 PLA=b1AIBBHb2
*LCRcb -
AIB13 PLA=b1AIBBHb2
*LCRllb 13
BA1 PLA=b0+b1BA 13
BA2 PLA=b1BAb2
9, 13
BA3 PLA=b1BAb2
*CL 13
BA4 PLA=b1BAb2
*mLCRb3
13
BA5 LN(PLA)=b0+b1LN(DBH)+b2LN(CL) -
CL1 LN(PLA)=b0+b1LN(CL) 10, 13
CL2 PLA=b0+b1CL+b2DBH -
CL3 PLA=b1CLb2
*DBH -
CV1 PLA=b1CVb2
13
SA1 PLA=b0+b1SABH 2, 7, 10, 12, 13
SA2 PLA=b0+b1SALLB 2, 7, 10, 12, 13
SA3 PLA=b0+b1SABH+b2CL 7, 10, 12, 13
SA4 PLA=b0+b1SALLB+b2CL -
SA5 PLA=b1SABHb2
*CL 3, 5, 10, 13
SA6 PLA=b1SABHb2
11, 13
SA7 PLA=b1SALLBb2
11, 13
SA8 LN(PLA)=b0+b1LN(SABH) 4, 10, 13
SA9 LN(PLA)=b0+b1LN(SALLB) 4, 10, 13
SA10 LN(PLA)=b0+b1LN(SABH)+b2LN(CL) 10, 13
SA11 PLA=b0+b1SABH+b2LCRCB 6, 13
SA12 PLA=b0+b1SABH+b2LCRLLB -
* Sources: 1, Shinozaki et al. 1964; 2, Marchand 1984; 3, Dean and Long 1986; 4,
Espinoza et al. 1987; 5, Long and Smith 1989; 6, Thompson 1989; 7, Coyea and
Margolis 1992; 8, Valentine et al.1994; 9, O’Hara and Valappil 1995; 10, Gilmore et al.
1996; 11, Maguire 1998; 12, Kenefic and Seymour 1999, 13, Meyer 2005.
Note: AIB, area inside bark; AIBBH, area inside bark at breast height; AIBLLB, area
inside bark at the lowest live branch; BA, basal area outside bark; CL, crown length; CV,
crown volume (0.33(CL x crown projection area)) (Seymour and Kenefic 1999); DBH,
diameter outside bark at breast height; LCRCB, live crown ratio at the base of the live

67. 53
crown; LCRLLB, live crown ratio at the lowest live branch; mLCR, modified live crown
ratio (ratio of CL to distance from leader to breast height) (Valentine et al. 1994); SA,
sapwood area; SABH, sapwood area at breast height; SALLB, sapwood area at the lowest
live branch.
of heteroscedasticity in the error variances. Log bias was determined for each of the log-
transformed models and a correction factor was calculated for each (after Snowdon
1991). Snowdon’s (1991) correction factor is the ratio of the arithmetic mean of the
dependent variable sample to the mean of the back-transformed predicted dependent
variables. Models were compared by their generalized r2
, FI score, and residual analysis
for both dependent variables.
Honer’s (1967) volume equation and parameter estimates were evaluated for
model bias against the observed stem volume from the 25 destructively sampled trees in
this study. The model fit was:
[3] V = DBH2
/ (4.167 + (244.906/H))
where V is the total stem volume inside bark, DBH is the outside bark diameter at breast
height in inches at 4.5 feet off the ground, and H is the total tree height in feet. Model [3]
was refit to the data to generate new parameter coefficients.
Mean annual volume increment (VINC, dm3
) for the two most recent complete
growth years was computed for the stem-analyzed trees in WinStem. Two nonlinear
models were fit to describe the relationship of VINC with respect to PLA and CFM:
[4] VINC = β1 x β2
[4a] VINC = β1
⎟⎟
⎟
⎠
⎞
⎜⎜
⎜
⎝
⎛
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟⎟
⎠
⎞
⎜⎜
⎝
⎛
−
3β
2β
x
-EXP1

68. 54
where VINC is the mean annual volume increment (dm3
) for the 2004 and 2005 growth
years, x is the independent variable PLA (m2
) or CFM (kg), and βi are regression
coefficients. Three weights (x-1
, x-2
, x-3
) were considered for each model.
Growth efficiency (GE) was quantified as VINC per unit leaf area (GEPLA,
dm3
/m2
) or VINC per unit crown foliage mass (GECFM, dm3
/kg). Both relationships
assume that crown attributes did not differ markedly between 2004 and the sampling
date. ANOVA was used to test for differences in GE among light exposure classes, soil
site classes, and presence of central decay. A significance level of α=0.10 was used due
to the low sample size in each categorical variable level.
Stem volume was estimated for the 296 northern white-cedar trees that were cored
(and described in Hofmeyer (2008), Ch.2) with model [3] and refit parameter
coefficients. Diameter outside bark and total height were back-grown two complete years
plus the partial ring that occurred during the season of sampling. Diameter increment was
analyzed in WinDendro and a constant proportion of tree-specific bark thickness to
diameter was assumed to estimate volume. Several regression models were screened to
predict height growth by diameter, height, live crown ratio, light exposure class, and/or
site class from the stem-analyzed stems; all resulted in nonsignificant parameters.
Therefore, the average annual height increment for the top meter of the destructively
sampled stems (0.08 m) was used to back-grow height two years in the cored tree sample.
Volume was estimated for the back-grown cored trees with model [3] with refit
regression coefficients. VINC was determined by subtracting the back-grown volume
from the most recent full year’s volume and dividing by 2 to get annual VINC.

69. 55
PLA and CFM were estimated for 296 cored northern white-cedar trees using
model [BA4]. This model had a low FI, a high r2
, and little bias in the residual analysis
for both dependent variables. Following the same protocol as above, models [4] and [4a]
were fit to describe the relationship between VINC and PLA or CFM. Log-linear VINC
models were screened for inclusion of age and site class into nonlinear models [4] and
[4a]. Stepwise log-linear least-squares regression rejected the inclusion of age (p=0.197)
and indicated that the VINC to PLA relationship was not appreciably affected by site
class (p=0.144). A second degree polynomial (Seymour and Kenefic 2002, Eq. 3) was
screened for influences of age and PLA on GE; both the age and the PLA coefficients did
not differ from 0. GEPLA and GECFM were quantified as the estimated VINC over the
estimated PLA and CFM for each cored tree. ANOVA was used to test for differences in
GE among light exposure classes, soil site classes, and presence of central decay at
α=0.05 using SYSTAT version 12. Models with multiple categorical independent
variables were tested for interactions among variable terms.
3.4 RESULTS
3.4.1 Branch Leaf Area and Foliage Mass
SLA ranged from 31.8 to 79.4 cm2
/g and differed by foliage location within the
crown (p<0.001). SLA was not different among site classes (p=0.465) or LECs (p=0.108)
(Table 3.4). The best-fit BFM and PLA model was [BLA 2] due to its low FI value, high
r2
, and normally distributed residuals (Table 3.5). This model had little bias in predicting
BLA from RD and branch diameter (Figure 3.2). For a given branch diameter, model [2a]
predicts a maximum BLA at 0.75 RD (Figure 3.3). Interestingly, this peak does not

70. 56
coincide with the maximum BFM predicted; BFM is estimated to peak at 0.25 RD.
Foliage peaks are more pronounced as branch diameter increases.
Table 3.4. Specific leaf area (cm2
/g) with respect to location within the crown, among site
classes, and among light exposure classes.
Location within crown N Mean SLA Standard error
Lower Quartile 25 61.7a
1.75
Mid Quartile 25 55.6b
1.75
Upper Half 25 46.0c
1.75
Site class
Organic 6 56.9 2.56
5 7 54.1 2.37
4 4 57.0 3.14
3 4 53.1 3.14
2 4 50.2 3.14
Light exposure class
1 9 57.2 2.03
2 3 59.3 4.31
3 8 53.3 2.03
4 4 51.2 3.05
5 1 43.2 -
Note: means followed by different letters are different at the α=0.05 level of significance
(Tukey’s HSD mean separation).
Table 3.5. Model evaluation for branch leaf area (cm2
) and branch foliage mass (g).
Model FI r2
b0 b1 b2 b3 b4
Log
bias χ2
Branch leaf area models
2 3881 0.638 6.851 0.064 - - - 0.963 <0.001
2a 3104 0.753 - 5.917 0.868 0.205 1.168 n/a 0.415
2b 2946 0.747 3.873 1.808 0.627 -1.099 - 1.086 0.019
Branch foliage mass models
2 67.19 0.601 2.298 0.060 - - - 0.974 <0.001
2a 60.10 0.787 - 0.754 0.912 0.404 1.104 n/a 0.613
2b 76.71 0.780 -0.467 1.941 0.459 -1.411 - 1.058 0.049
Note: FI – Calculated index of fit for transformed models (after Furnival 1961), rMSE –
the untransformed root mean square error (for comparison to the FI values; the lower of
the two is preferred), Log bias calculated after Snowdon (1991), χ2
– a Chi-squared test
of residual distribution. Normally distributed residuals have values >0.05.

71. 57
Branch basal diameter (cm)
0 2 4 6
Branchleafarea(m2
)
0
1
2
3
4
5
Predicted
Observed
A B
Relative distance into crown
0.0 0.2 0.4 0.6 0.8 1.0
B
Figure 3.2. Observed and predicted branch leaf area as a function of branch diameter (A)
and relative distance of the branch into the crown (B) with model [2a].
0 20 40 60 80
0.0
0.2
0.4
0.6
0.8
1.0
0 40 80 120 160 200
1 cm
2 cm
3 cm
Relativedistanceintocrown
Branch leaf area (cm2
) Branch foliage mass (g)
Top
Base
A B
Figure 3.3. Branch leaf area (A) and foliage mass (B) as a function of relative distance
into the crown and branch diameter as predicted by model [2a].

72. 58
3.4.2 Projected Leaf Area and Crown Foliage Mass
Fit statistics for all PLA and CFM models are presented in Appendix B. The best-
fit PLA and CFM model screened was [BA4], a modified Valentine et al. (1994) function
with basal area outside bark and modified live crown ratio as independent variables
(Table 3.6). Model [BA4] had the lowest FI and highest r2
of the CFM models, though
model [SA3] had the lowest FI of the PLA models. Model [SA3] had a b3 coefficient not
different from zero, limiting its applicability as a predictive model. [BA4] had the least
bias comparing observed PLA and CFM values from the 25 stem-analyzed trees with
values estimated from their respective tree core data (Figure 3.4). The remaining models
had poorer consistency among the two datasets. PLA of the 256 cored trees estimated
with model [BA 4] ranged from 9.11 to 161.17 m2
(mean= 57.91 m2
, SE=1.49) and CFM
ranged from 1.62 to 36.18 kg (mean=12.05 kg, SE=0.34).
Table 3.6. Best-fit area inside bark (AIB), basal area outside bark (BA), crown length
(CL), and sapwood area (SA) model for estimating projected leaf area (PLA) and crown
foliage mass (CFM), ranked by FI values. Refer to Table 3.3 for model forms.
Model Weight FI r2
b0 b1 b2 b3
PLA
SA3 SAbh
-2
12.61 0.747 -17.677* 0.415 2.634* -
BA4 BA-2
12.65 0.797 - 326.540 0.548 0.982
AIB2 AIBllb
-1
13.77 0.705 22.488 900.87 - -
CL1 n/a 14.73 0.665 0.012* 0.895 - -
CFM
BA4 BA-2
2.33 0.849 - 77.688 0.593 1.062
SA3 SAbh
-2
2.47 0.790 -4.924 0.077 0.755 -
AIB2 AIBllb
-1
2.86 0.719 4.364 193.589 - -
CL3 unweighted 3.16 0.715 - 0.130 0.468 -
* denotes a parameter coefficient not different than 0
Note: Furnival (1967) index value for model [CL3] is the rMSE value from the
unweighted model. Generalized r2
values were calculated after Kvalseth (1985).

73. 59
Basal area outside bark (m
2
)
0.00 0.04 0.08 0.12 0.16 0.20
Crownfoliagemass(kg)
0
10
20
Projectedleafarea(m2
)
0
30
60
90
120
Branch sum
Core data
A
B
Figure 3.4. Projected leaf area and crown foliage mass values calculated by branch
summation for 25 stem-analyzed cedar trees and estimated from their respective tree core
data with model [BA 4].

74. 60
3.4.3 Stem Volume
Refitting Honer’s (1967) model to the observed volume data resulted in new
regression coefficients of a1=2.139 and b1=301.634 (Imperial units). The refit model had
a generalized r2
of 0.970 and a residual sum of 6.88. Observed volume of the stem-
analyzed trees had a mean of 0.512 m3
(range of 0.075 to 1.237 m3
) while predicted
volume from the new coefficients had a mean of 0.504 m3
(range of 0.075 to 1.232 m3
).
The refit equation captured stem volume variation across all observed diameters (Figure
3.5). Estimated mean volume for the cored trees was 0.511 m3
(range of 0.053 to 2.511
m3
).
3.4.4 Volume Increment and Growth Efficiency
Mean annual VINC was 8.44 dm3
(range of 3.83 to 15.46 dm3
, SE=0.748) in the
25 stem-analyzed samples and 8.79 dm3
in the 256 cored trees (range of 0.99 to 34.29
dm3
, SE=0.312). Screened VINC models suggest that the relationship between VINC and
PLA/CFM was nearly linear in both the stem-analyzed and cored trees. The β2 coefficient
in [4] was significant for the stem-analyzed trees, but was not different from 1 with the
cored tree data (Table 3.7). [4a] models resulted in β3 coefficients that were not different
from 1 in any screened model, suggesting that the relationship between VINC and
PLA/CFM was not sigmoid. Estimated VINC from stem-analyzed and cored data is quite
similar over the range of PLAs observed in this study (Figure 3.6). Models were weighted
to reduce the influence of heteroscedacity, which resulted in a better fit with higher r2
values and lower FI values.

75. 61
DBH (cm)
10 20 30 40 50
Stemwoodvolume(dm3
)
0
300
600
900
1200
1500
Predicted volume
Observed volume
Figure 3.5. Observed stemwood volume analyzed in WinDendro with the estimated
volume from Honer’s (1967) equation with refit parameter coefficients.
Table 3.7. Parameters and fit statistics of the nonlinear regressions of VINC on PLA and
CFM for stem-analyzed and cored trees.
Model
Optimal
weight rMSE FI r2
b1 b2 b3
Stem-analyzed sample data
4 PLA-2
2.142 1.722 0.747 0.284 0.842*
CFM-2
2.186 1.841 0.727 1.205 0.796
4a PLA-2
1.957 1.744 0.752 19.850 92.015 1.115*
CFM-2
2.025 1.876 0.729 25.327 28.593 0.969*
Cored tree sample data
4 PLA-1
3.669 3.293 0.559 0.149 0.990*
CFM-2
3.699 3.282 0.560 0.886 0.925*
4a unweighted 3.660 n/a 0.541 30.770 131.873 1.308*
unweighted 3.659 n/a 3.659 30.739 29.073 1.209*
* Denotes parameter coefficients not different than 1 at the 95% confidence level.
Note: Model forms are described in text; rMSE, root mean square error of the unweighted
model; FI, Furnival’s index.

76. 62
Cored
Cored
Projected leaf area (m2
)
0 30 60 90 120 150 180
Annualstemwoodvolumeincrement(m2
)
0
10
20
30
40
Stem-analyzed
Stem-analyzed
Figure 3.6. Comparison of monotonic decreasing regression model [4] relating annual
stemwood volume increment to projected leaf area for 25 stem-analyzed and 256 cored
northern white-cedar trees.
ANOVA detected no differences in GEPLA and GECFM by site class or light
exposure class in the stem-analyzed trees (Table 3.8). No differences were detected in
GEPLA or GECFM among decayed and sound sample trees (p=0.24 and p=0.13,
respectively). No interaction was observed in the ANOVA model with the decay*site
term (p=0.465). Growth efficiency showed no trend with respect to increasing PLA or
CFM in the cored trees (Figure 3.7). ANOVA detected no differences in GEPLA or
GECFM by site or LEC in the cored trees (Table 3.9). GEPLA and GECFM were not
different among decay classes (presence/absence) (p=0.377 and p=0.283, respectively).

77. 63
Table 3.8. ANOVA results for growth efficiency of 25 destructively sampled northern
white-cedar trees by site and light exposure class.
Site class N Mean GEPLA* SE Mean GECFM SE
2 4 0.120 0.017 0.568 0.090
3 4 0.141 0.017 0.694 0.090
4 4 0.157 0.017 0.794 0.090
5 7 0.169 0.013 0.838 0.068
Organic 6 0.161 0.014 0.772 0.073
p-value 0.245 0.208
Light exposure
class
1 9 0.157 0.013 0.782 0.067
2 2 0.159 0.027 0.723 0.142
3 9 0.149 0.013 0.742 0.067
4 4 0.156 0.019 0.750 0.101
5 1 0.119 - 0.563 -
p-value 0.893 0.884
* GEPLA is stemwood volume increment per unit leaf area (dm3
/m2
), GECFM is
stemwood volume increment per unit crown foliage mass (dm3
/kg).

78. 64
Projected leaf area (m2)
0 30 60 90 120 150 180
Growthefficiency(dm3/m2)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
LEC 1
LEC 2
LEC 3
LEC 4
LEC5
A
Crown foliage mass (kg)
0 10 20 30 40
Growthefficiency(dm3/kg)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
B
Figure 3.7. Growth efficiency as a function of projected leaf area (A) and crown foliage
mass (B) by light exposure class in 296 cored northern white-cedar trees. Large symbols
identify the class mean.

79. 65
Table 3.9. ANOVA results for growth efficiency of 296 cored northern white-cedar trees
by site and light exposure class.
Site class N Mean GEPLA* SE Mean GECFM SE
2 32 0.164 0.012 0.794 0.059
3 40 0.152 0.011 0.745 0.053
4 50 0.133 0.009 0.653 0.047
5 104 0.156 0.006 0.764 0.033
Organic 70 0.157 0.008 0.774 0.040
p-value 0.201 0.264
Light exposure
class
1 65 0.137 0.008 0.700 0.042
2 88 0.152 0.007 0.752 0.036
3 92 0.153 0.007 0.774 0.035
4 13 0.171 0.011 0.817 0.056
5 14 0.173 0.018 0.819 0.090
p-value 0.116 0.482
* GEPLA is stemwood volume increment per unit leaf area (dm3
/m2
), GECFM is
stemwood volume increment per unit crown foliage mass (dm3
/kg).
3.5 DISCUSSION
3.5.1 Leaf Area Prediction
Mean SLA in this study (54.4 cm2
/g) suggests northern white-cedar foliar sprays
have more surface area per unit mass than red spruce (44.9 cm2
/g, Maguire et al. 1998) or
balsam fir (32.7 cm2
/g, Gilmore and Zenner 2005), but less than eastern hemlock (58.4
cm2
/g, Kenefic and Seymour 1999). SLA patterns indicate a high degree of plasticity in
foliar morphology. Northern white-cedar shade foliage in the lower crown tends to be
thin with a high surface area per unit mass, whereas sun foliage in the upper half of the
crown has a thicker cuticle and lower surface area per unit mass. Higher SLA values in
shade foliage often describe morphological responses to low light conditions, lower
temperatures, and less moisture stress than sun foliage (Barnes et al. 1998).

80. 66
Projected leaf area and foliage mass models may have been somewhat imprecise
given the high level of crown variability in the sample trees. Even the best-fit model
accounted for less than 80% of the variation in BLA and BFM. Model [2a] predicted a
peak in leaf area slightly below the midpoint of the crown, though branch leaf area was
relatively constant below 0.4 RD. Meyer (2005) used model [2a] to predict leaf area in
thinned and unthinned balsam fir and red spruce stands. Balsam fir branch leaf area
peaked at a RD of 0.6 (slightly below the crown midpoint) while red spruce peaked near
the crown top (0.25 RD). Gilmore and Seymour (1997) predicted peaks in balsam fir leaf
area at RD of 0.4, 0.55, 0.65, and 0.7 in balsam fir intermediate, suppressed, codominant,
and open-grown trees, respectively. This pattern suggests that balsam fir trees have leaf
area concentrations closer to the crown base in codominant and open-grown trees, while
those of inferior canopy positions have leaf area concentrations higher in the crowns.
Upper canopy northern white-cedar branch leaf area distribution appears to follow a
pattern more similar to balsam fir than red spruce.
In addition to a strong fit to the BLA summation data, model [BA 4] has the
benefit of requiring easily measured variables (dbh and LCR) to predict leaf area and
crown foliage mass. This allows for cost effective and efficient data collection, without
the need for coring trees in future northern white-cedar research.
3.5.2 Volume Increment and Growth Efficiency
Refitting Honer’s (1967) volume equation increased estimated volume by an
average 2.31 dm3
per tree (range of 0.097 to 7.08 dm3
), a 10.0% mean volume increase
(range of 3.4 to 14.7%). Annual VINC estimated from breast height cores with the refit
Honer’s (1967) equation was comparable to observed VINC measured using WinStem

81. 67
software. This suggests that two increment cores extracted in the field produced unbiased
estimates of stemwood volume increment.
The three-parameter Weibull function had nonsignificant regression coefficients
when fit to both the stem-analyzed and cored data which suggests that the relationship
between VINC and leaf area/foliage mass is not sigmoid. The β2 coefficient in the 2-
parameter power function (test of curvature) was not different from 1, indicating a nearly
linear relationship between VINC and PLA/CFM. A β2 coefficient equal to 1 has been
observed in understory strata in multiaged stands of shade-intolerant Pinus ponderosa
(O’Hara 1996) and Pinus contorta (Kollenberg and O’Hara 1999) and is indicative of
constant VINC across all observed leaf areas. This relationship is atypical for trees
classified as tolerant of shade. Though there are accounts of northern white-cedar as a
midtolerant species (Curtis 1946), most classify it as shade-tolerant (Scott and Murphy
1987, Johnston 1990) because it can replace itself in nearly pure stands. Though our
observed VINC/PLA pattern observed was unexpected for a shade-tolerant species, it
must be stressed that no overtopped trees were sampled and their relationship is
unknown.
A scatter plot with linear least squares regression of GE as a function of breast
height age (both corrected and uncorrected for PLA effects) has a slope not different than
0 and an r2
of 0.01 (Figure 3.8). Seymour and Kenefic (2002) found that age was not
significant in predicting GE in red spruce or eastern hemlock until the prediction model
adjusted for a peaking GE at intermediate PLAs. Screening their model in the present
study resulted in b1, b2, and b3 coefficients not different than zero. This result supports the
notion that inclusion of an age term in GE is not warranted for northern white-cedar in

82. 68
Breast height ring count
0 50 100 150 200 250
Growthefficiency(dm
3
/m
2
)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Figure 3.8. Growth efficiency as a function of breast height age.
mixed-species multicohort stands. Sound trees with breast height ages ranging from 35 to
235 years were included in this study; however, there appears to be no GE reduction in
trees approaching 250 years of age.
Because GE is an ecophysiological measure of growing space occupancy,
understanding and manipulation of associated environmental variables is important to
silviculture. The apparent lack of GE response to site class, LEC, decay, and age is
particularly important given the growth habit of northern white-cedar in mixed-species
stands. It is a tree commonly found in subordinate crown positions with its highest
abundance on poorly drained mineral and organic sites with approximately 80% of the
outwardly sound trees centrally decayed. There are opportunities to favor large-crowned
cedar trees, regardless of current canopy position and soil drainage. Though mean PLA
was higher in trees of superior canopy position, PLA had a wide range within all LECs

83. 69
(see Figure 3.7). These characteristics, in combination with the apparent lack of age
response, suggest that northern white-cedar could be managed successfully with
multiaged silviculture.
Northern white-cedar has been described as a plastic tree species with highly
variable morphology and growth strategies depending upon its immediate environment
(Briand et al. 1992, Matthes-Sears and Larson 1995). One attribute of northern white-
cedar that has been noted is its “evenness” of growth regardless of site conditions
(Harlow 1927) or stand density (Foltz and Johnston 1968). Results from this study
provide support for northern white-cedar’s ability to maintain similar VINC per unit PLA
with respect to several environmental variables. This species possesses GE patterns
conventionally ascribed to intolerant tree species, yet it is commonly found in
subordinate canopy positions. Perhaps it is meaningful to describe northern white-cedar
not by its shade tolerance, but as a stress-tolerant tree species (after Grime 1977) in that it
can withstand moisture- and nutrient-limited environments such as cliff faces (Matthes-
Sears et al. 1995), is often in subordinate canopy positions in mixed-species stands,
readily adjusts osmotic regulation to withstand drought conditions (Collier et al. 1989),
can act as both a colonizing pioneer tree species and a climax species, and responds well
to release after extended periods of suppression (Hofmeyer 2008, Ch. 4). Constant GE of
a given PLA with respect to site and light variables may be an additional indication of its
stress tolerance and plasticity.
It must be emphasized that the results of this study pertain to northern white-cedar
trees that were outwardly sound. Though no significant GE differences were attributed to
the presence of central decay, outward signs of defect are common in this species.

84. 70
Testing only growth capabilities of outwardly sound individuals may represent a “best-
case” scenario. Though no differences in GE were detected among LECs, low GE values
have been associated with overtopped trees, particularly in multiaged stands (Seymour
and Kenefic 2002). Further research into pure stands of even- and multi-aged northern
white-cedar is recommended to investigate the influence of stand structure and leaf area
allocation to stand-level volume growth.
Unlike many previous growth efficiency studies, it was not possible to select trees
with perfect crown architecture to estimate leaf area, foliage mass, and stem volume.
Northern white-cedar has weak epinastic control which frequently results in main stem
forking (Curtis 1946). Nearly 80% of the trees cored prior to selecting the sound stem-
analysis sample trees were decayed. Tree selection thus included some stems with small
forks and some trees with small volumes of central decay. Though this does not negate
GE results, it may have lead to PLA and VINC model imprecision relative to previous
studies. However, it is likely that the results from this study are applicable to outwardly
sound northern white-cedar trees throughout northern Maine.
3.6 CONCLUSIONS
Prediction of leaf area and stem volume requires easily obtained tree variables for
use in future northern white-cedar research. The relationship of volume increment to tree-
level projected leaf area and foliage mass was found to be nearly linear, not sigmoid as
has been observed in associated shade-tolerant conifers. Analysis suggests that growth
efficiency was unaffected by site class, light exposure, presence of central decay, and tree
age. This study documented the adequacy of estimating crown variables and volume
increment with both stem-analyzed and cored trees. Results indicate that there are

85. 71
opportunities for increasing tree-level volume increment by managing for large crowned
individuals across site gradients. This study bolsters ecological literature suggesting that
northern white-cedar trees are adaptable to a wide variety of environmental conditions
and that its successional niche remains somewhat of an anomaly.

86. 72
Chapter 4:
HISTORICAL EARLY STEM DEVELOPMENT OF NORTHERN WHITE-
CEDAR IN MAINE
4.1 ABSTRACT
Evidence from 80 stem-analyzed northern white-cedar trees > 17.4 cm dbh
indicates that early height and diameter growth were slow; approximately 42 years were
required to reach sapling size and 96 years to reach pole size. Mean annual height growth
was 0.08 m through the first 4 m of height development. Approximately 80% of the
sample trees had initial growth suppression followed by a growth release, suggesting a
history as advance regeneration. Mean period of initial suppression exceeded 60 years,
and some individuals responded to release after nearly 200 years of suppression. Growth
releases frequently occurred during or slightly after known spruce budworm epidemics.
History of advance regeneration and response to release suggests that this species might
respond well to uneven-aged and shelterwood silvicultural systems. Foresters are
recommended to encourage advance regeneration and avoid residual understory damage
during partial harvest operations.

87. 73
4.2 INTRODUCTION
The forests of northern Maine lie within the Acadian Forest complex, a
transitional forest between the boreal forest of eastern Canada and the northern
hardwood-hemlock forests of northeastern United States (Braun 1950). In Maine, the
most abundant coniferous tree species by timberland area and growing stock volume are
balsam fir (Abies balsamea (L.) Mill), red spruce (Picea rubens Sarg.), and northern
white-cedar (Thuja occidentalis L.) (McWilliams et al. 2005). Early accounts of forest
history in Maine were primarily focused on eastern white pine (Pinus strobus L.), red
spruce, and balsam fir. Red spruce and balsam fir were often the focus of regeneration
and pulpwood volume studies because of their prevalence, importance in spruce
budworm (SBW, Choristoneura fumiferana) dynamics, and commercial value (e.g.
Westveld 1931).
Red spruce and balsam fir are susceptible to episodic SBW outbreaks; the
budworm is a defoliating insect with a natural cycle of 30 to 80 years (Royama 1984).
Forest stands dominated by mature balsam fir tend to be completely killed in severe
outbreaks, while mixed-species stands with a higher component of red spruce are more
robust due to red spruce’s lower vulnerability (Osawa et al. 1986). Dendrochronology
studies suggest that little, if any, growth reduction is observed in non-host species such as
northern white-cedar (Krause 1997, Fraver et al. 2007). Northern white-cedar was
commonly a residual in mixed-species stands following episodic SBW outbreaks (Fraver
et al. 2007).
Extensive partial cuttings that left balsam fir residuals and the SBW outbreak ca.
1913-1919 resulted in depletion of merchantable softwoods in Maine (Seymour 1992).

88. 74
Removal of the overstory during this era released advance spruce and fir regeneration and
resulted in predominately even- or two-aged stands. A SBW epidemic from the mid-
1970s to the late 1980s infested dense spruce-fir stands that had been released during the
previous outbreak. Clearcutting became a common harvesting practice to salvage or pre-
salvage fir-dominated stands (Seymour 1992). Because no advance regeneration had
formed under these dense even-aged stands (many still in the stem exclusion phase),
harvesting operations regenerated stands that were truly even-aged.
Northern white-cedar has a lesser known role in Maine’s forest history, though
accounts from surveyor records from the early 1800s indicate that it was common in
northern Maine forests (Lorimer 1977). Northern white-cedar trees with good stem form
were exploitatively harvested for poles during the early 1900s, primarily from dense
lowland stands (Cogbill 1985, Fraver 2004). A better understanding of northern white-
cedar history, ecology, and silviculture is needed to sustainably provide critical winter
habitat for white-tailed deer (Odocoileus virginianus) and niche commodity products
such as shingles, fence posts, and mulch. Unfortunately, northern white-cedar has been
historically under- represented in the forestry literature throughout its native range
(Hofmeyer et al. 2007), often leading to uninformed management decisions.
Recent findings from the U.S. Forest Service, Forest Inventory and Analysis
(FIA) suggest that sustainability of the northern white-cedar resource in Maine might be a
concern. Data from McWilliams et al. (2005) indicate that between 1982 and 2003,
northern white-cedar forestland declined from approximately 417,000 hectares to
388,000 hectares (Figure 4.1a). Growing stock volume increased from 1982 to 1995
(approximately 52 million m3
to 56 million m3
), but decreased from 1995 to 2003 (48

89. 75
Timberland(thousandsofhectares)
0
100
200
300
400
500
Sawtimber
Poletimber
Seedling/sapling
1982 1995 2003
Growingstock(millionsofm3
)
0
10
20
30
40
50
A
B
Figure 4.1. Timberland area (A) and growing stock volume (B) of northern white-cedar
by stand size class in Maine (after McWilliams et al. 2005).

90. 76
million m3
), though sawtimber growing stock volumes increased from 1995 to 2003
(Figure 4.1b).
It appears that existing northern white-cedar trees are getting larger, but that there
is low ingrowth and recruitment into the poletimber class. Negative net change in
northern white-cedar growing stock volume was most attributable to cull increment and
harvest levels that exceeded net growth. Mean annual mortality for all dominant tree
species in Maine was 1.2%; northern white-cedar annual mortality rate was
approximately 0.6%. Ingrowth and accretion combined (gross growth) far exceeded
mortality, confirming that excessive mortality was not the likely cause of growing stock
declines.
The apparent lack of northern white-cedar ingrowth and recruitment is
disconcerting given slow early height and diameter growth rates reported for this species
in Maine (Curtis 1946), Michigan (Nelson 1951), Minnesota (Cornett et al. 2001),
Wisconsin (Rooney et al. 2002), and Quebec (LaRouche 2007). Northern white-cedar
regeneration abundance and growth rates have been shown to be negatively correlated
with deer abundance because of its palatability and slow height growth rates (Nelson
1951, Van Deelen 1999, Cornett et al. 2000). As a result of high deer densities and slow
height development in many regions throughout its range, northern white-cedar stands
have a strong component of competing species in the regeneration stratum, often leading
to projected species composition shifts (Thornton 1957, Johnston 1972, Scott and
Murphy 1986, Cornett et al. 2000). On the Big Reed Forest Preserve in northern Maine,
Fraver (2004) noted abundant northern white-cedar regeneration but a distinct lack of
sapling and poletimber recruitment since the early 1900s. Anecdotal discussions with

91. 77
foresters in Maine suggest that northern white-cedar stands frequently have sparse cedar
regeneration and abundant seedlings and saplings of competing species such as balsam
fir.
Understanding early stem development of northern white-cedar trees is imperative
to sustainable management of the northern white-cedar resource given recent statewide
inventory data. The objective of this study was to identify historical height and diameter
growth patterns from reconstruction of sound northern white-cedar trees from throughout
northern Maine.
4.3 METHODS
One hundred northern white-cedar stems were provided by Maibec Industries,
Inc. in St-Pamphile, Quebec, Canada in September, 2005. All stems were harvested from
northern Maine from known stand locations during March, 2005 (Figure 4.2). Aside from
location, neither stand-level data nor standing tree-level data were available for these
samples. Of the 100 stems initially provided, 59 were usable for reconstruction of early
stem height and diameter growth (i.e. sound to the pith). Four cross-sectional discs were
taken from each stem; one at stump height (SH, 0.3 m), just above breast height (BH, 2.0
m), mid height (MH, 4.2 m), and top height (TH, variable heights). Cross-sectional discs
were taken at these intervals to allow the stems to be processed in a cedar shingle mill
after sampling. Stems that had been bucked to remove sections of decay prior to mill
transport were excluded. A subset of 21 sound cedar trees that had been stem-analyzed
and used to determine northern white-cedar leaf area (LA) and growth efficiency
(Hofmeyer 2008, Ch. 3) were also included in this analysis. LA trees were sectioned at 1-
m intervals from 0.3 m to top height. Stand-level and tree-level information was available

92. 78
Figure 4.2. Site locations of the destructively sampled northern white-cedar stems from
Maibec Industries (♦) and the growth efficiency study (■).

93. 79
for these trees. All sites were mixed-species stands, though stand density and the
proportion of northern white-cedar, red spruce, balsam fir, and hardwoods varied by site
(Table 4.1).
Cross-sectional discs were dried, sanded, and analyzed in Regent Instruments
WinDendro at 1200 dpi resolution. A portion of annual radial growth was estimated for
two stem-analyzed trees due to small incidences of central decay; each decayed section
was less than 2.5 cm in diameter at SH. A pith locator consisting of a clear transparency
Table 4.1. Stand-level characteristics for 21 stem-analyzed northern white-cedar trees.
Plot basal area (m2
/ha)
Tree DBH Site Class Cedar Spruce-fir Hardwood Total Mean DBH
6 15.9 Organic 62.0 2.3 0.0 64.3 28.5
61 33.5 4 6.9 9.2 9.2 32.1 27.1
63 24.2 4 6.9 9.2 9.2 32.1 27.1
71 29.3 Organic 39.0 2.3 4.6 52.8 29.3
73 36.8 Organic 39.0 2.3 4.6 52.8 29.3
161 24.8 Organic 25.3 13.8 2.3 43.6 28.6
165 29.2 Organic 25.3 13.8 2.3 43.6 28.6
176 35.2 5 23.0 11.5 0.0 39.0 30.0
180 41.5 5 23.0 11.5 0.0 39.0 30.0
186 22 5 20.7 16.1 2.3 39.0 21.8
187 13.9 5 20.7 16.1 2.3 39.0 21.8
188 32.1 5 20.7 16.1 2.3 39.0 21.8
211 20.6 5 18.4 9.2 0.0 27.5 20.8
215 25.9 5 18.4 9.2 0.0 27.5 20.8
408 32.6 2 9.2 16.1 9.2 34.4 25.1
576 19.1 4 11.5 6.9 4.6 23.0 46.7
578 24.7 4 11.5 6.9 4.6 23.0 46.7
589 42.7 5 39.0 11.5 2.3 52.8 43.8
590 23 5 39.0 11.5 2.3 52.8 43.8
598 23.7 3 18.4 27.5 2.3 48.2 37.9
600 29.3 3 18.4 27.5 2.3 48.2 37.9
Note: Site class follows Brigg’s (1994) descriptions; spruce-fir is the sum of red spruce
and balsam fir basal area; mean DBH is the mean diameter of all trees tallied on the site.

94. 80
with concentric circles of constant radial increment was used to estimate missing growth
years (Applequist 1958). Estimated values were within the minimum and maximum
observed annual growth values for all sampled trees. To analyze both sample groups
concurrently, the number of annual rings per meter of height growth was standardized for
all trees. For the remainder of the study, SH will refer to 0.3 m, BH to 2.0 m and MH to
4.2 m. Because the top height disc was highly variable in height and diameter,
comparison among sample trees at that point is not meaningful.
Number of years required to reach 2.5, 12.7, 22.9, and 38.1 cm inside bark at SH
as well as number of years required to reach 2.5, 12.7, and 22.9 cm at BH and MH were
determined for each sample tree. All year values reported reflect the time required for
trees to reach a given diameter at a given height starting from 0.3 m, not from ground
level. SH diameters ranged from 17.4 to 55.0 cm (mean=40.3, SE=0.82) and SH age
ranged from 87 to 356 years (mean=183.6, SE=5.73).
A running 10-year radial increment average was used to identify periods of
growth release, wherein mean radial growth of ten years after an event is subtracted from
the mean radial growth of ten years prior to an event. Fraver and White (2005) suggested
an absolute increase threshold of 0.41 mm to identify release events in northern white-
cedar. Using an absolute increase, instead of a relative percentage increase (Nowacki and
Abrams 1997), reduces the likelihood of identifying false positive releases arising from
low mean growth values. Early sapling growth was evaluated to determine if sample
stems originated in a gap or in suppression (after Lorimer et al. 1988). Because most
stems in this study are from unknown stands, probability measures of origin were not

95. 81
employed. A gap origin threshold of 1 mm for the mean of the first 5 years of growth at
BH for all trees was used in this study.
4.4 RESULTS
Mean annual height growth was 0.083 m over the first 4 m, suggesting that on
average one could expect seedlings to grow from SH to 1 m in approximately 15.5 years
and to 4 m in approximately 50.0 years (Table 4.2). Early diameter growth was also slow.
On average, it took 42 years for cedar in this study to reach sapling size (2.5 cm at BH),
95 years to reach poletimber size (12.7 cm at BH), 140 years to reach sawtimber size
(22.9 cm at BH), and 170 years to reach shingle stock size (38.1 cm at SH) (Table 4.3).
Several growth patterns were common in the sample trees. A period of
suppression followed by a growth release was observed in 64 SH samples (80% of
sample trees) (e.g. Figure 4.2); initial suppression was observed in only 15 MH samples
(19%). In addition, nearly every sampled individual exhibited prominent basal flaring
(Figure 4.3), often times extending well above breast height, as well as constant or
slightly increasing annual radial increment (increasing area increment) over time at BH
(see Appendix C).
Though this was designed to be a tree reconstruction study, not a stand reconstruction
study, some notable release patterns were observed at BH (see Appendix C). Of the 80
stems sampled, only 13 (16%) had early radial increments high enough to be considered
of gap origin. All others had radial increment suggesting a period of initial suppression.
Only 18 sample trees failed to show signs of release in their tree ring series; i.e. no mean
annual increment change exceeding the 0.41 mm mean threshold. Trees not of gap origin
that were released had a period of initial suppression ranging from 13 to 197 years

96. 82
(mean=65.9, SE=6.61); release periods frequently coincided with historical SBW
epidemics (Figure 4.5).
Table 4.2. Mean annual height increment (m) of 80 stem-analyzed northern white-cedar
trees.
Mean annual height increment (m)
Stem height (m) minimum mean maximum standard error
1.0 0.027 0.084 0.250 0.006
2.0 0.028 0.082 0.283 0.006
3.0 0.034 0.081 0.208 0.004
4.0 0.036 0.083 0.195 0.003
Number of years required to reach a given height*
1.0 4 15.50 36 0.893
2.0 6 26.70 61 1.464
3.0 13 38.42 75 1.658
4.0 19 49.97 103 1.915
* Number of years to reach a given height from stump height (0.3 m).
Table 4.3. Number of years required to reach a given inside bark diameter at a given
height for 80 stem-analyzed northern white-cedar trees.
Height* Diameter n min mean max SE
SH 2.5 80 6 26.4 65 1.43
SH 12.7 80 27 85.5 177 3.18
SH 22.9 80 47 120.1 257 3.93
SH 38.1 66 81 170.1 317 5.67
BH 2.5 80 9 42.0 86 2.02
BH 12.7 80 28 96.0 171 3.35
BH 22.9 80 54 139.9 238 4.35
MH 2.5 80 22 65.1 126 2.39
MH 12.7 80 42 111.9 198 3.39
MH 22.9 64 81 166.3 284 4.93
* SH, BH, and MH refer to 0.3, 2.0, and 4.2 m respectively. Diameter units are cm.

97. 83
0
2.5
5
1890 1910 1930 1950 1970 1990 2010
Radialincrement(mm)
0
2.5
5
1890 1910 1930 1950 1970 1990 2010
Radialincrement(mm)
Tree 600 at 4.3 m
Tree 600 at 0.3 m
0
2.5
5
1890 1910 1930 1950 1970 1990 2010
Radialincrement(mm)
0
2.5
5
1890 1910 1930 1950 1970 1990 2010
Radialincrement(mm)
Tree 600 at 4.3 m
Tree 600 at 0.3 m
Figure 4.3. Typical pattern of suppression followed by a release and relatively constant
radial growth in the stump height disc (top) and no signs of a suppressed core at the mid
height disc (bottom).

98. 84
Tree 42, 114 years at stump
2
4
6
8
10
12
15 30 4545 30 15
Diameter (cm)
Height(m)
Tree 42, 114 years at stump
2
4
6
8
10
12
15 30 4545 30 15
Diameter (cm)
Height(m)
2
4
6
8
10
12
15 30 4545 30 15 15 30 4545 30 15
Diameter (cm)
Height(m)
Figure 4.4. Stem profile of sample tree M 42. Each line represents one year of height and
diameter growth. Closely spaced lines are periods of reduced radial and height increment.
Note the prominent basal flare below 2 m and the suppressed core.

99. 85
M 37 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 36
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 176
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 64
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure 4.5. Breast height ring chronologies of four northern white-cedar trees in Maine.
M37G was of gap origin; M36 had a suppressed origin and was never released; vertical
dashed lines indicate growth releases in M64 and LA176, coinciding with known spruce
budworm outbreaks.

100. 86
4.5 DISCUSSION
This study provides evidence that many mature northern white-cedar trees from
the working Maine landscape had prolonged periods of slow growth. Mean SH age of
sawtimber trees was nearly 140 years and mean age of shingle stock trees exceeded 170
years. These values far exceed the typical lifespan of 80 years suggested by Larson
(1991) for non-cliff dwelling northern white-cedar trees. In addition, even the oldest
sample trees in the present study had relatively constant radial increment, suggesting that
age-induced radial growth senescence does not occur. Lowland northern white-cedar
trees have a reported maximum lifespan of 400 to 500 years in the Lake States and
commonly live longer than associates in mixed-species stands (Johnston 1990, Pregitzer
1990); the oldest sample in the present study was 356 years at SH.
Sampling only sound individuals might bias the results of this study. However,
because approximately 80% of the outwardly sound northern white-cedar trees in Maine
are centrally decayed (Hofmeyer 2008, Ch.2), there is little alternative for reconstructing
early growth. Early growth rates in cedar trees with internal decay remain unknown;
however, mean annual height increment values found in this study are consistent with
findings observed in cedar seedlings and saplings in the Lake States (0.05 m/yr in
suppression) (Heitzman et al. 1997). Hannah (2004) found height growth to range from
0.15 m/yr to 0.3 m/yr depending upon site conditions in young even-aged stands in
Vermont, suggesting approximately 40 years to reach 6 m. It is difficult to tease apart
causes for slow height development from reconstructed stems with unknown stand
histories.

101. 87
Several associated shade-tolerant conifer species exhibit a similar period of initial
height and diameter growth suppression when regeneration occurred under an enclosed
canopy. Both red spruce and balsam fir are shade-tolerant and can become established
and survive under the shade of larger trees (Frank 1990, Blum 1990). It has been argued
that northern white-cedar is also a shade-tolerant conifer due to its ability to successfully
regenerate and replace itself (Scott and Murphy 1987, Johnston 1990).
Northern white-cedar is a preferred browse species for white-tailed deer in
northern climates (Ullrey et al. 1964, 1967, 1968). Excessive seedling herbivory has been
reported as a cause for a lack of northern white-cedar sapling recruitment in the Lake
States (Van Deelen 1999, Cornett et al. 2000, Rooney et al. 2002). Slow early height
development, even under the best circumstances observed here, required northern white-
cedar seedlings to be within browsing height (<2 m) for nearly a decade. It is likely that
deer browse does not fully explain the initial period of growth suppression, as white-
tailed deer were not present in northern Maine from European settlement to the 1890s
(Stanton 1963) and many of these samples originated during that time.
The majority of stem-analyzed cedar trees in this study likely originated as
advance regeneration. Common patterns of early growth suppression followed by a
release support the notion that northern white-cedar can regenerate under shade, but may
require a significant canopy opening to be recruited into the sapling or pole size classes.
Though some sample trees responded with a growth release within 15 years of reaching
breast height, it was most common for trees to respond to releases after >50 years of
suppression. Four sample trees responded well after over 150 suppressed years. It appears
that cedar trees can persist for well over 300 years without ever experiencing a growth

102. 88
release (see Appendix C). In a study of stem-analyzed cedar trees from seven cedar-
dominated stands in Michigan, Heitzman et al. (1997) reported a tripling of annual height
growth in previously established saplings following significant disturbances. They
reported growth responses to release after approximately 135 years of suppression.
Many trees in the present study that did respond to release maintained constant or
increasing radial increment. This radial increment phenomenon was reported previously
in the literature by Harlow (1927), who noted trees from two northern white-cedar
communities in New York had remarkable “evenness” in their annual growth rings. Trees
generally allocate photosynthate hierarchically first to maintenance respiration of living
tissue, production of fine roots and leaves, reproductive tissues, primary growth, and
finally to xylem increment and defensive compounds (Oliver and Larson 1996). Trees
that allocate a constant proportion of photosynthate to stemwood growth over time
(constant area increment) have a pattern of decreasing radial increment, as would be
expected from geometry of a circle. As trees age and grow larger, maintenance
respiration requirements increase often resulting in reduced area increment over time.
Advanced tree age has also been associated with reductions in photosynthesis per unit
foliage biomass (Day et al. 2001). Trees in subordinate canopy positions may also have a
lower proportion of photosynthate available for stemwood growth after higher priority
requirements have been satisfied (Assman 1970, Oliver and Larson 1996). Given that
many of the sample trees were of advanced age and northern white-cedar commonly
occupies subordinate canopy positions, patterns of increasing area increment at stump
and breast height are difficult to explain. Pressler’s “Law” suggests that below the base of
the live crown, annual area increment should be constant, corresponding to tapering

103. 89
radial increment from the base of the live crown to the base of the bole (with a slight
increase near stump height) (e.g. Figure 4.6). Very few northern white-cedar trees
sampled in this study have annual area increment consistent with that expectation
(Appendix A.4). Because northern white-cedar has a low specific gravity, is weak and
brittle in compression (Seeley 2007), and is commonly associated with poor sites where
soil and root stability is poor, northern white-cedar trees may have developed a strategy
to tolerate that stress through reductions in height:diameter ratio by preferentially
allocating resources to lower portions of the stem. Thus, patterns of increasing area
increment observed in these trees might not necessarily be a sign of increasing vigor, but
rather a growth strategy of supporting increasingly larger trees on a weak structural base.
In a detailed reconstruction of the Big Reed Forest Preserve in northern Maine,
Fraver (2004) suggested that of all the species sampled, northern white-cedar was the
most complacent (i.e. least responsive to local environmental conditions). Though
northern white cedar may be complacent relative to other tree species, findings from this
study suggest that northern white-cedar responds strongly with increased radial increment
to favorable changes in the environment. In fact, Heitzman et al. (1997) suggested that
most cedar trees in their study sites were established and/or released in response to stand-
level disturbance resulting from historical logging practices.
Unfortunately, there could be no stand reconstructions in this study. Breast height
chronologies provide historical evidence to suggest that northern white-cedar trees may
have experienced growth releases during or slightly following SBW epidemics. Known
historical SBW outbreaks in northern Maine occurred ca. 1810-1813, 1913-1919, and
1972-1986 (Krause 1997, Fraver et al. 2007). Though not every tree shows release during

104. 90
Annual growth
0 1 2 3 4 5
DiscHeight(m)
0
2
4
6
8
10
12
14
Area increment (cm2)
Radial increment (mm)
Figure 4.6. Mean annual area and radial increment for Tree LA 6 consistent with
Pressler’s Law. The horizontal dashed line indicates the base of the live crown.
each of the SBW epidemics that it survived, each of the known SBW epidemics since the
early 1800s coincided with a growth release in at least one of the sampled cedar trees.
Because SBW is a defoliating insect that reduces leaf area of its primary hosts (balsam fir
and spruce species), growth increases would be expected of non-host species in mixed-
species stands.
Though stand-level data were not available for the Maibec sample trees, they
were for the LA trees. Of the 21 LA sample trees, 10 trees experienced a growth release

105. 91
coinciding with the 1913-1919 SBW epidemic and 9 had a release coinciding with the
1972-1986 SBW epidemic. Fraver (2004) found lowland dense northern white-cedar
stands had the least response during SBW outbreaks. Interestingly, Tree LA 6 in this
study experienced a growth release coinciding with the 1970’s SBW outbreak though
only 4% of its stand basal area was SBW host species and >60 m2
/ha of non-host
northern white-cedar basal area. All LA trees that had multiple releases in their
chronology were from sites with >25% of current basal area in SBW host species.
Because there was no stand reconstruction, past abundance of host and non-host species
is unknown. It is probable that stands with a high proportion of northern white-cedar
basal area show little growth increase during SBW epidemics relative to those occupying
mixed-species stands with a higher proportion of host species.
Regardless of the cause of release, it is clear that suppressed northern white-cedar
trees can respond to localized environmental changes, even after extended periods of
suppression. This finding has important implications for northern white-cedar
silviculture. Historical evidence of advance regeneration encourages the use of uneven-
aged silviculture or variants of the shelterwood system, with options for long-lived
reserve trees. Strong responses in height and diameter increment suggest that thinning
even-aged stands could be a successful intermediate treatment to increase growth on
stems with good form. Foresters are encouraged to promote advance regeneration during
partial harvest entries without targeting overtopped suppressed cedar trees for removal.
Because of the weak and brittle wood properties of northern white-cedar, caution to avoid
residual stand damage should be taken. Protect understory cedar saplings regardless of
age if cedar is a desired overstory species in future stands.

106. 92
4.6 CONCLUSIONS
Mean annual height increment of northern white-cedar trees in this study was 0.08
m through the first 4 m of development. Mean annual diameter increment was also slow;
requiring a mean of 42 years to reach the sapling class (maximum 86 years), 96 years to
poletimber (maximum 177 years), 140 years to sawtimber (maximum 257 years), and 170
years to sawtimber (maximum 317 years). Approximately 80% of the sample trees had
growth releases after extended periods in suppression, often with constant or increasing
radial increment beyond that time. This historical early stem development pattern
suggests the importance of past advance regeneration. Because suppressed northern
white-cedar trees can respond to release, there are options for uneven-aged or
shelterwood treatments in cedar stands. Avoiding residual stand damage to saplings and
seedlings is strongly encouraged to foster advance regeneration and recruitment into
sapling and pole classes.
Future research in mixed-species stands should investigate responses of northern
white-cedar as a non-host of the SBW given the coincidental findings in this study.
Contemporary regeneration and recruitment data for northern white-cedar in Maine is
critically needed to inform management decisions.

107. 93
LITERATURE CITED
Aldous, S.E. 1941. Deer management suggestions for northern white cedar types. Journal
of Wildlife Management 5(1): 90-94.
Applequist, M. B. 1958. A simple pith locator for using with off-center increment cores.
Journal of Forestry 56: 141.
Archambault, S. and Y. Bergeron. 1992. Discovery of a living 900-year-old northern
white cedar, Thuja occidentalis, in northwestern Quebec. Canadian Field Naturalist
106(2): 192-195.
Assman, E. 1970. The Principles of Forest Yield Study. Transl. By S.H. Gardiner.
Oxford: Pergamon Press Ltd.
Avery, T.E. and H.E. Burkhart. 1994. Forest Measurements, 4th ed. Chapter 14. McGraw
Hill Publishing, New York, NY.
Barnes, B.V., D.R. Zak, S.R. Denton, and S.H. Spurr. 1998. Forest Ecology, 4th ed. Ch.
8. John Wiley and Sons. New York, NY.
Basham, J.T., P.V. Moork, and A.G. Davidson. 1953. New information concerning
balsam fir decays in eastern North America. Canadian Journal of Botany 31:334-357.
Bechtold, W.A. 2003. Crown position and light exposure classification - An alternative to
field-assigned crown class. North. J. Appl. For. 20(4):154-160.
Behr, E.A. 1974. Distinguishing heartwood in northern white cedar. Wood Science 6(4):
394-395.
Binkley, D. and P. Reid. 1984. Long-term responses of stem growth and leaf area to
thinning and fertilization in a Douglas-fir plantation. Can. J. For. Res. 14: 656-660.
Blum,B.M. 1990. Picea rubens Sarg. – red spruce. In: Silvics of North America. Vol. 1.
Conifers. U.S. Department of Agriculture Agricultural Handbook 654. 20 p.
Braun, E. L. 1950. Deciduous forests of eastern North America. Haher Press, New York,
NY. 596 p.
Briand, C.H., U. Posluszny, and D.W. Larson. 1991. Patterns of architectural variation in
Thuja occidentalis L. (eastern white cedar) from upland and lowland sites. Botanical
Gazette 152(4): 494-499.
Briand, C.H., U. Posluszny, and D.W. Larson. 1992. Differential axis architecture in
Thuja occidentalis (eastern white cedar). Canadian Journal of Botany 70(2): 340-348.

108. 94
Briand, C.H., U. Posluszny, and D.W. Larson. 1993. Influence of age and growth rate on
radial anatomy of annual rings of Thuja occidentalis L. (eastern white cedar).
International Journal of Plant Sciences 154(3): 406-411.
Briggs, R.D. 1994. Site classification field guide. ME Ag. For. Exp. Sta. Misc. Publ. 724.
15 p.
Briggs, R.D. and R.C. Lemin, Jr. 1994. Soil drainage class effects on early response of
balsam fir to precommercial thinning. Soil Science Society of America Journal 58:
1231-1239.
Brix, H. and A.K. Mitchell. 1983. Thinning and nitrogen fertilization effects on sapwood
development and relationships of foliage quantity to sapwood area and basal area in
Douglas-fir. Canadian Journal of Forest Research 13: 384-389.
Buckley, B.M., R.J.S. Wilson, P.E. Kelly, D.W. Larson, and E.R. Cook. 2004. Inferred
summer precipitation for southern Ontario back to AD 610, as reconstructed from
ring widths of Thuja occidentalis. Canadian Journal of Forest Research 34: 2541-
2553.
Carmean, W.H., J.T. Hahn, and R.D. Jacobs. 1989. Site index curves for forest tree
species in the eastern United States. USDA Forest Service General Technical Report
NC-128. 142 p.
Cary, A. 1896. Third annual report of the Forest Commissioner of the State of Maine,
Augusta, Maine, USA.
Caulkins, H.L., Jr. 1967. The ecology and reproduction of northern white-cedar. M.S.
Thesis. University of Michigan. 70 p.
Chimner, R.A. and J.B. Hart. 1996. Hydrology and microtopography effects on northern
white-cedar regeneration in Michigan's Upper Peninsula. Canadian Journal of Forest
Research 26(3): 389-393.
Collier, D.E. and M.G. Boyer. 1989. The water relations of Thuja occidentalis L. from
two sites of contrasting moisture availability. Botanical Gazette 150(4): 445-448.
Cornett, M.W., K.J. Puettmann, L.E. Frelich, P.B. Reich. 2001. Comparing the
importance of seedbed and canopy type in the restoration of upland Thuja
occidentalis forests in northeastern Minnesota. Restoration Ecology 9(4): 386-396.
Cornett, M.W., P.B. Reich, and K.J. Puetmann. 1997. Canopy feedbacks and
microtopography regulate conifer seedling distribution in two Minnesota conifer-
deciduous forests. Ecoscience 4(3): 353-364.
Cornett, M.W., P.B. Reich, K.J. Puettmann, and L.F. Frelich. 2000. Seedbed and
moisture availability determine safe sites for early Thuja occidentalis (Cupressaceae)
regeneration. American Journal of Botany 87(12): 1807-1814.

109. 95
Coyea, M.R. and Margolis, H.A. 1992. Factors affecting the relationship between
sapwood area and leaf area of balsam fir. Canadian Journal of Forest Resources 22:
1684-1693.
Curtis, J.D. 1944. Northern white cedar on upland soils in Maine. Journal of Forestry 42:
756-759.
Curtis, J.D. 1946. Preliminary observations on northern white cedar in Maine. Journal of
Ecology 27(1): 23-36.
Davis, A., K. Puettmann, and D. Perala. 1998. Site preparation treatments and browse
protection affect establishment and growth of northern white-cedar. Research Paper
NC-330, U.S. Department of Agriculture, Forest Service, North Central Experiment
Station. 9 p.
de Blois, S. and A. Bouchard. 1995. Dynamics of Thuja occidentalis in an agricultural
landscape of southern Quebec. Journal of Vegetation Sciences 6(4): 531-542.
Dean, T.J. and Long, J.N. 1986. Variation in sapwood area – leaf area relations within
two stands of lodgepole pine. Forest Science 32: 749-758.
Edwards, D.R. and M.A. Dixon. 1995. Mechanisms of drought response in Thuja
occidentalis L. I. Water stress conditioning and osmotic adjustment. Tree Physiology
15(2): 121-127.
Espinosa Bancalari, M.A., Perry, D.A. and Marshall J.D. 1987. Leaf are – sapwood area
relationships in adjacent young Douglas-fir stands with different early growth rates.
Canadian Journal of Forest Research 15: 113-120.
Fernald, M.L. 1919. Lithological factors limiting the ranges of Pinus banksiana and
Thuja occidentalis. Rhodora 21: 41-67.
Fernandez, I.J. 1992. Characterization of eastern U.S. spruce-fir soils. In: Eager, C. and
M.B. (eds.) Ecology and Decline of Red Spruce in the Eastern United States.
Springer-Verlag, NY. pp. 40-63.
Foltz, B.W. and W.F. Johnston. 1968. Gross basal area growth of northern white-cedar is
independent of stand density over a wide range. Research Note NC-61. U.S.
Department of Agriculture, Forest Service, North Central Experiment Station. 4 p.
Frank, R.M. 1990. Abies balsamea (L.) Mill. – balsam fir. In: Silvics of North America.
Vol. 1. Conifers. U.S. Department of Agriculture Agricultural Handbook 654. 22 p.
Fraver, S. 2004. Spatial and temporal patterns of natural disturbance in old-growth forests
of northern Maine, USA. Ph.D. Dissertation. University of Maine. Orono, ME. 185 p.
Fraver, S. and A.S. White. 2005. Identifying growth releases in dendrochronological
studies of forest disturbance. Canadian Joural of Forest Research 35: 1648-1656.

110. 96
Fraver, S., R.S. Seymour, J.H. Speer, and A.S. White. 2007. Dendrochronological
reconstruction of spruce budworm outbreaks in northern Maine, USA. Canadian
Journal of Forest Research 37: 523-529.
Furnival, G.M. 1961. An index for comparing equations used in constructing volume
tables. Forest Science 7: 337-341.
Gevorkiantz, S.R. and W.A. Duerr. 1939. Volume and yield of northern white cedar in
the Lake States: a progress report. U.S. Department of Agriculture, Forest Service,
Lake States Forest Experiment Station. 55 p.
Gilmore, D.W. and E.K. Zenner. 2005. Foliage-sapwood area relationships for balsam fir
in north-central Minnesota. Northern Journal of Applied Forestry 2(3): 203-210.
Gilmore, D.W. and R.S. Seymour. 1996. Alternative measures of stem growth efficiency
applied to Abies balsamea from four canopy positions in central Meaine, USA. Forest
Ecology and Management 84: 209-218.
Gilmore, D.W. and R.S. Seymour. 1997. Crown architecture of Abies balsamea from four
canopy positions. Tree Physiology 17: 71-80.
Gilmore, D.W., R.S. Seymour, and D.A. Maguire. 1996. Foliage-sapwood area
relationships for Abies balsamea in central Maine, U.S.A. Can. J. For. Res. 26: 2071-
2079.
Godman, R.M. 1958. Silvical characteristics of northern white-cedar (Thuja
occidentalis). Department of Agriculture, United States Forest Service Lake States
Forest Research Experiment Station Paper Number 67. 17p.
Grier, C.C. and R.H. Waring. 1974. Conifer foliage mass related to sapwood area. Forest
Science 20: 205-206.
Grime, J.P. 1977. Evidence for the existence of three primary strategies in plants and its
relevance to ecological and evolutionary theory. The American Naturalist 111(982):
1169-1194.
Habeck, 1958. White cedar ecotypes in Wisconsin. Ecology 39: 457-463.
Hannah, P.R. 2004. Stand structures and height growth patterns in northern white-cedar
stands on wet sites in Vermont. Northern Journal of Applied Forestry 21(4): 173-179.
Harlow, W.M. 1927. The effect of site on the structure and growth of white cedar Thuja
occidentalis L. Ecology 8(4): 453-470.
Hart, A.C. 1963. Spruce-fir silviculture in northern New England. In: Proceedings,
Society of American Foresters. Boston, MA. 1963. pp. 107-110.

111. 97
Heitzman, E., K.S. Pregitzer, and R.O. Miller. 1997. Origin and early development of
northern white-cedar stands in northern Michigan. Canadian Journal of Forest
Research 27: 1953-1961.
Heitzman, E., K.S. Pregitzer, R.O. Miller, M. Lanasa, and M. Zuidema. 1999.
Establishment and development of northern white-cedar following strip clearcutting.
Forest Ecology and Management 123: 97-104.
Hepting, G.H. 1971. Diseases of Forest and Shade Trees of the United States. USDA For.
Serv. Ag. Handbook Num. 386. p. 478-480.
Hofmeyer, P.V., L.S. Kenefic, and R.S. Seymour. 2007. Northern white-cedar: An
annotated bibliography. Cooperative Forestry Research Unit Research Report 07-01.
30 p.
Holcombe, J.W. 1976. The bryophyte flora of Thuja seedbed logs in a northern white-
cedar swamp. The Michigan Botanist 15: 173-181.
Honer, T.G. 1967. Standard volume tables and merchantable conversion factors for the
commercial tree species of central and eastern Canada. Forest Management Research
and Services Institute, Ottawa, Canada.
Innes, J.C., M.J. Ducey, J.H. Gove, W.B. Leak, and J.P. Barrett. Size-density metrics,
leaf area, and productivity in eastern white pine. Canadian Journal of Forest Research
35: 2469-2478.
Irland, L.C. Wildlands and Woodlots: The Story of New England's Forests. University
Press of New England. 217 p.
Johnston, W.F. 1977. Manager’s handbook for northern white cedar in the North Central
States. General Technical Report NC-35. U.S. Department of Agriculture, Forest
Service, North Central Forest Experiment Station. 18 p.
Johnston, W.F. 1990. Thuja occidentalis L. – northern white-cedar. In: Silvics of North
America. Vol. 1. Conifers. U.S. Department of Agriculture Agricultural Handbook
654: 1189-1209.
Jokela, E.J. and T.A. Martin. 2000. Effects of ontogeny and soil nutrient supply on
production, allocation, and leaf area efficiency in loblolly and slash pine stands.
Canadian Journal of Forest Research 30: 1511-1524.
Jokela, J.J. and C.L. Cyr. 1977. Performance of northern white-cedar in central Illinois.
In: Proceedings of the 13th Lake States Forest Tree Improvement Conference.
General Technical Report NC-50. U.S. Department of Agriculture, Forest Service,
North Central Forest Experiment Station. pp. 100-106.
Kangas, P. 1989. Comparison of two northern white cedar (Thuja) forests. The Michigan
Botanist 28(2): 59-66.

112. 98
Kelly, P.E. and D.W. Larson. 1997. Dendroecological analysis of the population
dynamics of an old-growth forest on cliff-faces of the Niagara Escarpment, Canada.
Journal of Ecology 85: 467-478.
Kelly, P.E., E.R. Cook, and D.W. Larson. 1992. Constrained growth, cambial mortality,
and dendrochronology of ancient Thuja occidentalis on cliffs of the Niagara
Escarpment: an eastern version of bristlecone pine? International Journal of Plant
Sciences 153: 117-127.
Kelly, P.E., E.R. Cook, and D.W. Larson. 1994. A 1397-year tree-ring chronology of
Thuja occidentalis from cliff faces of the Niagara Escarpment, southern Ontario,
Canada. Canadian Journal of Forest Research 24(5): 1049-1057.
Kenefic, L.S. and R.S. Seymour. 1999. Leaf area prediction models for Tsuga canadensis
in Maine. Canadian Journal of Forest Research 29: 1574-1582.
Kollenberg, C.L. and K.L. O'Hara. Leaf area and tree increment dynamics of even-aged
and multiaged lodgepole pine stands in Montana. Canadian Journal of Forest
Research 29: 687-695.
Krause, C. The use of dendrochronological material from buildings to get information
about past spruce budworm outbreaks. Canadian Journal of Forest Research 27: 69-
75.
Kvålseth, T.O. 1985. Cautionary note about R2
. Am. Stat. 39: 279-285.
Lamy, S., A. Bouchard, and J.P. Simon. 1999. Genetic structure, variability, and mating
system in eastern white cedar (Thuja occidentalis) populations of recent origin in an
agricultural landscape in southern Quebec. Canadian Journal of Forest Research
29(9): 1383-1392.
Larouche, C., J-C. Ruel, J-M. Lussier, and L.S. Kenefic. 2006. Regeneration of Thuja
occidentalis L. in mixedwood stands on mesic sites after partial cuts in Quebec. In:
Proceedings of the Eastern Canada – USA Forest Science Conference. 19-21 Oct.
2006, Quebec, QC: 90-94.
Larouche, C., S. Morissette, J-C. Ruel, J-M. Lussier, and L.S. Kenefic. 2007.
Regeneration of Thuja occidentalis L. in mixedwood stands after partial cutting. In:
Proceedings of the Carrefour de la recherché forestière. 19-20 Sept. 2007, Quebec,
QC.
Larson, D.W. 2001. The paradox of great longevity in a short-lived tree species.
Experimental Gerentology 36(4): 651-673.
Larson, D.W. and P.E. Kelly. 1991. The extent of old-growth Thuja occidentalis on cliffs
of the Niagara Escarpment. Canadian Journal of Botany 69: 1628-1636.

113. 99
Larson, D.W., J. Doubt, U. Matthes-Sears. 1994. Radially sectored hydraulic pathways in
the xylem of Thuja occidentalis as revealed by the use of dyes. International Journal
of Plant Sciences 155(5): 569-582.
Larson, D.W., U. Matthes-Sears, and P.E. Kelly. 1993. Cambial dieback and partial shoot
mortality in cliff-face Thuja occidentalis: evidence for sectored radial architecture.
International Journal of Plant Sciences 154: 496-505.
Little, E.L., Jr., 1971, Atlas of United States trees, volume 1, conifers and important
hardwoods: U.S. Department of Agriculture Miscellaneous Publication 1146, 9 p.,
200 maps.
Loehle, C. 1996. Optimal defense investment in plants. Oikos 75: 299-302.
Long, J.N. and F.W. Smith. 1990. Determinants of stemwood production in Pinus
contorta var. latifolia forests: the influence of site quality and stand structure. Journal
of Applied Ecology 27: 847-856.
Lorimer, C. G. 1977. The presettlement forest and natural disturbance cycle of
northeastern Maine. Ecology 58: 139-148.
Lorimer, C.G., L.E. Frelich, and E.V. Nordheim. 1988. Estimating gap origin
probabilities for canopy trees. Ecology 69: 778-785.
MacLean, D.A. and W.E. MacKinnon. 1997. Effects of stand and site characteristics on
susceptibility and vulnerability of balsam fir and spruce to spruce budworm in New
Brunswick. Canadian Journal of Forest Research 27: 1859-1871.
Maguire, D.A. and Bennett, W.S. 1996. Patterns in vertical distribution of foliage in
young coastal Douglas-fir. Canadian Journal of Forest Research 26: 1991-2005.
Maguire, D.A., J.C. Brissette, and L. Gu. 1998. Crown structure and growth efficiency of
red spruce in uneven-aged, mixed-species stands in Maine. Can. J. For. Res. 28:
1233-1240.
Manion, P.D. 1991. Tree Disease Concepts. Second Ed. Prentice Hall. 416 p.
Marchand, P. 1984. Sapwood area as an estimator of foliage biomass and projected leaf
area for Abies balsamea and Picea rubens. Canadian Journal of Forest Research 14:
85-87.
Matthes-Sears, U, C.H. Nash, and D.W. Larson. 1995. Constrained growth of trees in a
hostile environment: the role of water and nutrient availability for Thuja occidentalis
on cliff faces. International Journal of Plant Sciences 156(3): 311-319.
Matthes-Sears, U. and D.W. Larson. 1991. Growth and physiology of Thuja occidentalis
L. from cliffs and swamps: is variation habitat or site specific? Botanical Gazette
152(4): 500-508.

114. 100
Matthes-Sears, U., C. Neeser, and D.W. Larson. 1992. Mycorrhizal colonization and
macronutrient status of cliff-edge Thuja occidentalis and Acer saccharum. Ecography
15(3): 262-266.
Matthes-Sears, U., P.E. Kelly, C.E. Ryan, and D.W. Larson. 2002. The formation and
possible ecological function of stem strips in Thuja occidentalis. International Journal
of Plant Sciences 163(6): 949-958.
Matthes-Sears, U., S.C. Stewart, and D.W. Larson. 1991. Sources of allozymic variation
in Thuja occidentalis in southern Ontario, Canada. Silvae Genetica 40(3/4): 100-105.
McWilliams, W.H.; Butler, B.J., Caldwell, L.E., Griffith, D.M., Hoppus, M.L., Laustsen,
K.M., Lister, A.J., Lister, T.W., Metzler, J.W., Morin, R.S., Sader, S.A., Stewart,
L.B., Steinman, J.R., Westfall, J.A., Williams, D.A., Whitman, A., Woodall, C.W.
2005. The forests of Maine: 2003. USDA, Forest Service Resource Bulletin NE-164.
188 p.
Meng, X. and R.S. Seymour. 1992. Influence of soil drainage on early development and
biomass production of young, herbicide-released fir-spruce stands in north central
Maine. Canadian Journal of Forest Research 22: 955-967.
Meyer, S.R. 2005. Leaf area as a growth predictor of Abies balsamea and Picea rubens in
managed stands in Maine. M.S. Thesis. University of Maine. Orono, ME. 117 p.
Miller, R.O. 1990. Guidelines for establishing animal exclosures for research in cedar
stands. CEDAR Action Group Note No. 1. 7 p.
Miller, R.O., D. Elsing, M. Lanasa, and M. Zuidema. 1990. Northern white-cedar: stand
assessment and management options. In: Proceedings of the Northern White Cedar in
Michigan Workshop. D.O. Lantagne, Ed. Michigan State University Agricultural
Experimental Station Research Report 512. pp. 47-56.
Murphy, L. S. 1917. The red spruce: its growth and management. United States
Department of Agriculture Bulletin 544.
Musselman, R.C., D.T. Lester, and M.S. Adams. 1975. Localized ecotypes of Thuja
occidentalis L. in Wisconsin. Ecology 56: 647-655.
Nelson, T.C. 1951. A reproduction study of northern white cedar including results of
investigations under federal aid in wildlife restoration project Michigan 49-R.
Department of Conservation Game Division. Lansing, Michigan. 100 p.
Nichols, N.S., T.G. Gregoire, and S.M. Zedaker. 1991. The reliability of tree crown
position classification. Canadian Journal of Forest Research 21: 698-701.
Nowacki, G.J., and M.D. Abrams. 1997. Radial-growth averaging criteria for
reconstructing disturbance histories from presettlement-origin oaks. Ecological
Monographs 67: 225-249.

115. 101
O’Hara, K.L. 1988. Stand structure and growing space efficiency following thinning in
an even-aged Douglas-fir stand. Canadian Journal of Forest Research 18: 859-866.
O’Hara, K.L. and Valappil, N.I. 1995. Sapwood-leaf area prediction equations for multi-
aged ponderosa pine stands in western Montana and central Oregon. Canadian
Journal of Forest Research 25: 1553-1557.
O'Hara, K.L. 1996. Dynamics and stocking-level relationships of multiaged ponderosa
pine stands. Forest Science Mongraph 33.
Osawa, A. 1989. Causality in mortality patterns of spruce trees during a spruce budworm
outbreak. Canadian Journal of Forest Research 19: 632-638.
Ostrofsky, W.D. and J. A. Dirkman. 1991. A survey of logging damage to residual timber
stands harvested for wood biomass in southern Maine. Maine Ag. Exp. Sta. Misc.
Rpt. 363. 8 p.
Ostrofsky, W.D., R.S. Seymour and R.C. Lemin, Jr. 1986. Damage to northern
hardwoods from thinning using whole-tree harvesting technology. Canadian Journal
of Forest Research 16: 1238-1244.
Pregitzer, K.S. 1990. The ecology of northern white cedar. Northern White Cedar in
Michigan Workshop. D.O. Lantagne, Ed. Michigan State University Agricultural
Experiment Station Research Report 512: 7-12.
Reiners, W.A. 1974. Foliage production by Thuja occidentalis L. from biomass and litter
fall estimates. American Midland Naturalist 92(2): 340-345.
Ringius, G.S. 1979. Status of eastern white cedar, Thuja occidentalis, in western Nova
Scotia. Canadian Field Naturalist 93(3): 326-328.
Roberts, S.D. and J.N. Long. 1992. Production efficiency of Abies lasiocarpa: influence
of vertical distribution of leaf area. Can. J. For. Res. 22: 1230-1234.
Roberts, S.D., J.N. Long, and F.W. Smith. 1993. Canopy stratification and leaf area
efficiency: a conceptualization. Forest Ecology and Management 60: 143-156.
Roe, E.I. 1947. Thinning in cedar swamps. USDA Forest Service Lake States Forest
Experiment Station Technical Note Number 279. 1 p.
Rooney, T.P., S.L. Solheim, and D.M. Waller. 2002. Factors affecting the regeneration of
northern white cedar in lowland forests of the Upper Great Lakes region, USA. Forest
Ecology and Management 163: 119-130.
Rosso, P. and E. Hansen. 1998. Tree vigour and the susceptibility of Douglas fir to
Armillaria root disease. European Journal of Forest Pathology 28: 43-52.

116. 102
Royama, T. 1984. Population dynamics of the eastern spruce budworm Choristioneura
fumiferana. Ecological Monograph 45: 429-462.
Sabine, D.L., W.B. Ballard, G. Forbes, J. Bowman, and H. Witlaw. 2001. Use of
mixedwood stands by wintering white-tailed deer in southern New Brunswick.
Forestry Chronicle 77(1): 97-103.
Scott, M.L. and P.G. Murphy. 1986. Origin and composition of an old-growth cedar-
hardwood stand: the role of dune activity. The Michigan Botanist 25(2): 57-65.
Scott, M.L. and P.G. Murphy. 1987. Regeneration patterns of northern white-cedar, an
old-growth forest dominant. American Midland Naturalist 117: 10-16.
Seely, O. 2007. Physical properties of common woods. Available on the Web at
http://www.csudh.edu/oliver/chemdata/woods.htm. Last Accessed 11/2007.
Seymour, R. S. 1992. The red spruce-balsam fir forest of Maine: evolution of silvicultural
practice in response to stand development patterns and disturbances. In: M. J. Kelty,
B. C. Larson, and C. D. Oliver, editors. The ecology and silviculture of mixed-species
forests. Kluwer Publishers, Norwell, Massachusetts, USA. pp. 217-244.
Seymour, R.S. and L.S. Kenefic. 2002. Influence of age on growth efficiency of Tsuga
canadensis and Picea rubens trees in mixed-species, multiaged northern conifer
stands. Canadian Journal of Forest Research 32: 2032-2042.
Seymour, R.S. and M.A. Fajvan. 2001. Infuence of prior growth suppression and soil on
red spruce site index. Northern Journal of Applied Forestry 18(2): 55-62.
Shigo, A.L. 1984. Compartmentalization: a conceptual framework for understanding how
trees grow and defend themselves. Annual Review of Phytopathology 22: 189-214.
Shinozaki, K., Yoda, K., Hozumi, K. and Kira, T. 1964b. A quantitative analysis of plant
form – the pipe model. II. Further evidence of the theory and its application in forest
ecology. Japanese Journal of Ecology 14: 133-139.
Smith, D.M., B.C. Larson, M.J. Kelty, and P.M.S. Ashton. 1997. The Practice of
Silviculture: Applied Forest Ecology. John Wiley & Sons. New York. 537 p.
Smith, F.W. and J.N. Long. 1989. The influence of canopy architecture on stemwood
production and growth efficiency of Pinus contorta var. latifolia. Journal of Applied
Ecology 26(2): 681-691.
Snowdon, P. 1991. A ratio estimator for bias correction in logarithmic regressions.
Canadian Journal of Forest Research 21: 720-724.
Stanton, D.C. 1963. A history of the white-tailed deer in Maine. Maine Department of
Inland Fisheries and Game. Game Division Bulletin No. 8. 75 p.

117. 103
Steinman, J.R. 1992. A comprehensive evaluation of spruce-fir growth and yield in
Maine as related to physical and chemical soil properties. Ph.D. Dissertation,
University of Maine. Orono, ME. 124 p.
Tardif, J. and Y. Bergeron. 1997. Comparative dendroclimatological analysis of two
black ash and two white cedar populations from contrasting sites in the Lake
Duparquet region, northwestern Quebec. Canadian Journal of Forest Research 27(1):
108-116.
Thompson, D.C. 1989. The effect of stand structure and stand density on the leaf area-
sapwood area relationship of lodgepole pine. Can. J. For. Res. 19: 392-396.
Thornton, P.L. 1957a. Problems of managing Upper Michigan’s coniferous swamps.
Journal of Forestry 55: 192-197.
Thornton, P.L. 1957b. Effects of cutting methods on advance reproduction in two mature
mixed coniferous swamps in Upper Michigan. Journal of Forestry 55: 448-451.
Trimble, G.R. Jr. 1969. Diameter growth of individual hardwood trees. Research Paper
NE-145. U.S. Department of Agriculture, Forest Service, Northeastern Forest
Experiment Station. 25 p.
Ullrey, D.E., W.G. Youatt, H.E. Johnson, L.D. Fay, and B.E. Brent. 1967. Digestibility of
cedar and jack pine browse for the white-tailed deer. Journal of Wildlife Management
31(3): 448-454.
Ullrey, D.E., W.G. Youatt, H.E. Johnson, L.D. Fay, B.E. Brent, and K.E. Kemp. 1968.
Digestibility of cedar and balsam fir browse for the white-tailed deer. Journal of
Wildlife Management 32(1): 162-171.
Ullrey, D.E., W.G. Youatt, H.E. Johnson, P.K. Ku, and L.D. Fay. 1964. Digestibility of
cedar and aspen browse for the white-tailed deer. Journal of Wildlife Management
28(4): 791-797.
Valentine, H.T., Baldwin, V.C., Gregoire, T.C. and Burkhart, H.E. 1994. Surrogates for
foliar dry matter in loblolly pine. For. Sci. 40: 576-585.
Van Deelen, T.R. 1999. Deer-cedar interactions during a period of mild winters:
Implications for conservation of conifer swamp deeryards in the Great Lakes Region.
Natural Areas Journal 19: 263-274.
Van Deelen, T.R., K.S. Pregitzer, and J.B. Haufler. 1996. A comparison of presettlement
and present-day forests in two northern Michigan deer yards. American Midland
Naturalist 135(2): 181-194.

118. 104
Velazquez-Matrinez, A., D.A. Perry, and T.E. Bell. 1992. Response of aboveground
biomass increment, growth efficiency, and foliar nutrients to thinning, fertilization
and pruning in young Douglas-fir plantations in the central Oregon Cascades.
Canadian Journal of Forest Research 22: 1278-1289.
Verme, L.J. 1965. Swamp conifer deeryards in northern Michigan: their ecology and
management. Journal of Forestry 63: 523-529.
Vose, J.M. and H.L. Allen. 1988. Leaf area, stemwood growth, and nutrition
relationships in loblolly pine. Forest Science 34(3): 547-563.
Waring, R.H. 1987. Characteristics of trees predisposed to die. BioScience 37(8): 569-
574.
Waring, R.H., W.G. Thies, and D. Muscato. 1980. Stem growth per unit leaf area: a
measure of tree vigor. Forest Science 26: 112-117.
Westveld, M. 193 1. Reproduction on the pulpwood lands in the northeast. United States
Department of Agriculture, Technical Bulletin 223. 52 p.
Whitney, R.D. 1989. Root rot damage in naturally regenerated stands of spruce and
balsam fir in Ontario. Canadian Journal of Forest Research 19: 295-308.
Whitney, R.D. 1997. Relationship between decayed roots and aboveground decay in
three conifers in Ontario. Canadian Journal of Forest Research 27: 1217-1221.
Williams, R.A., B.F. Hofmman, and R.S. Seymour. 1990. Comparison of site index and
biomass production of spruce-fir stands by soil drainage class in Maine. Forest
Ecology and Management 41: 279-220.
Worrall, J.J., L.D. Thomas, T.C. Harrington. 2005. Forest dynamics and agents that
initiate and expand canopy gaps in Picea-Abies forests of Crawford Notch, New
Hampshire, USA. Journal of Ecology 93 (1), 178–190.
Zon, R. 1914. Balsam fir. United States Department of Agriculture, Bulletin 55. 68 p.

119. 105
APPENDICES

120. 106
Appendix A
SITE INDEX RING CHRONOLOGIES
BF 1
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 4
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 9
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 10
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 18
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1. Ring chronologies of the sample trees selected to quantify site index. The title
of each figure denotes the tree species (balsam fir (BF), red spruce (RS), or northern
white-cedar (NWC)) and the unique identification sample number.

121. 107
BF 23
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 24
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 30
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 37
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 41
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 43
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1.Continued.

122. 108
BF 45
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 48
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 49
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 50
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 53
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 55
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued

123. 109
NWC 57
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 60
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 61
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 64
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 70
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 81
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

124. 110
RS 83
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 88
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 91
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 95
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 98
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 99
0
1
2
3
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

125. 111
RS 112
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 121
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 124
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 126
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 140
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 141
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

126. 112
BF 147
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 148
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 160
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 165
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 169
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 173
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

127. 113
BF 181
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 182
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 185
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 194
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 195
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 198
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

128. 114
BF 199
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 200
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 206
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 224
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 236
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 267
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

129. 115
RS 286
0
2.5
5
7.5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 288
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 296
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 298
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 300
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 308
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

130. 116
BF 316
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 317
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 329
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 338
0
2.5
5
7.5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 340
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 365
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

131. 117
RS 373
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 381
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 384
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 389
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 393
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 394
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

132. 118
NWC 396
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 398
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 403
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 411
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 412
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 415
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

133. 119
BF 422
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 440
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 441
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 444
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 453
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 471
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

134. 120
RS 473
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 474
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 482
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 483
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 485
0
2.5
5
7.5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 494
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

135. 121
BF 501
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 502
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 510
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 511
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 512
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 523
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

136. 122
BF 541
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 544
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 545
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 554
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 555
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 562
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

137. 123
BF 573
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
RS 583
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
NWC 600
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 602
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 622
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 623
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

138. 124
BF 625
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
BF 634
0
2.5
5
1820 1860 1900 1940 1980 2020
Radialincrement(mm)
Figure A.1 Continued.

139. 125
Appendix B
FIT STATISTICS FOR PROJECTED LEAF AREA AND CROWN FOLIAGE
MASS REGRESSION MODELS
Table B.1. Fit statistics for PLA models.
Model Form weight rMSE FI r2
b0 b1 b2 b3
SA3 y=b0+b1*SAbh+b2*CL SAbh-2
0.120 12.651 0.747 -
17 677
0.415 2.634* -
BA4 y=b1*BAobb2
*mLCRb3
BA-2
197.859 12.683 0.797 - 326.540 0.548 0.982
SA12 y=b0+b1*SAbh+b2*LCRllb SAbh-2
0.128 13.590 0.709
-
18 694 0.493 29.667* -
SA5 y=b1*SAbhb2
*CL SAbh-2
0.129 13.664 0.699 - 0.427* 0.527 -
SA6 y=b1*SAbhb2
SAbh-2
0.129 13.687 0.691 - 0.255* 1.142 -
SA1 y=b0+b1*SAbh SAbh-2
0.130 13.729 0.689
-
6 534* 0.562 - -
BA5 LN(y)=b0+b1*LN(DBH)+b2*LN(CL) n/a 0.267 13.737 0.776
-
0 863* 0.761 0.513 -
AIB02 y=b0+b1*AIBllb
AIBllb-
1 79.606 13.774 0.705 22.488 900.870 - -
SA10 LN(y)=b0+b1*LN(SAbh)+b2*LN(CL) n/a 0.270 13.863 0.709 -1.119 0.627 0.485 -
SA11 y=b0+b1*SAbh+b2*LCRcb SAbh-2
0.131 13.875 0.697
-
10 081 0.512 17.714* -
CL1 LN(y)=b0+b1*LN(CL) n/a 0.287 14.731 0.665 0.012* 0.895 - -
SA9 LN(y)=b0+b1*LN(SAllb) n/a 0.287 14.731 0.665 0.012* 0.895 - -
AIB04 y=b1*AIBllbb2
- 14.758 n/a 0.715 - 352.160 0.539 -
SA4 y=b0+b1*SAllb+b2*CL - 14.811 n/a 0.701 -3.290 0.501 1.569* -
SA2 y=b0+b1*Sallb - 14.880 n/a 0.698 4.813* 0.595 - -
SA7 y=b1*SAllbb2
- 14.905 n/a 0.710 - 0.934* 0.921 -
SA8 LN(y)=b0+b1*LN(SAbh) n/a 0.292 15.025 0.658 -1.397 1.142 - -
BA3 y=b1*BAobb2
*CL BA-2
239.460 15.349 0.693 - 10.781 0.271 -
AIB06 y=b1*AIBbhb2
*mLCR - 15.353 n/a 0.692 - 276.893 0.444 -
AIB13 y=b1*AIBbhb2
*LCRllb - 15.523 n/a 0.685 - 282.166 0.419 -
AIB08 y=b1*AIBbhb2
*mLCRb3
- 15.586 n/a 0.696 - 277.877 0.462 0.823
BA2 y=b1*BAobb2
BA-1
247.970 15.718 0.665 - 310.829 0.642 -
AIB09 y=b1*AIBbhb2
*CL
AIBbh-
2 312.766 15.746 0.647 - 7.488 0.132 -
AIB12 y=b1*AIBbhb2
*LCRcb - 15.780 n/a 0.669 - 289.735 0.347 -
CL3 y=b1*CLb2
*dbhob - 16.211 n/a 0.657 - 0.903* 0.318* -
CL2 y=b0+b1*CL+b2*dbhob - 16.355 n/a 0.635
-
20 232 2.197* 1.830 -
BA1 y=b0+b1*BAob - 17.135 n/a 0.599 20.624 481.954 - -
AIB01 y=b0+b1*AIBbh
AIBbh-
1 78.108 17.526 0.564 24.693 519.734 - -
AIB03 y=b1*AIBbhb2
- 17.734 n/a 0.589 - 287.501 0.566 -
AIB07 y=b1*AIBbh*mLCRb2
- 21.311 n/a 0.624 - 881.225 0.296 -
AIB05 y=b1*AIBbh*mLCR - 21.988 n/a 0.654 - 1029.954 - -
AIB11 y=b1*AIBbh*LCRllb - 22.540 n/a 0.645 - 1108.187 - -
AIB10 y=b1*AIBbh*LCRcb - 24.671 n/a 0.654 - 1320.782 - -

140. 126
Table B.2. Fit statistics for CFM models.
Model Form weight rMSE FI r2
b0 b1 b2 b3
BA4 y=b1*BAobb2
*mLCRb3
BA-2
36.317 2.328 0.849 - 77.688 0.593 1.062
SA3 y=b0+b1*SAbh+b2*CL Sabh-2
0.023 2.465 0.790 -4.924 0.078 0.755 -
SA5 y=b1*SAbhb2
*CL SAbh-2
0.024 2.532 0.770 - 0.075* 0.564 -
SA4 y=b0+b1*SAllb+b2*CL Sallb-2
0.033 2.692 0.753 -3.436* 0.079* 0.799 -
BA3 y=b1*BAobb2
*CL BA-2
43.591 2.794 0.773 - 2.587 0.326 -
SA12 y=b0+b1*SAbh+b2*LCRllb SAbh-2
0.027 2.841 0.721 -5.162 0.100 8.373* -
AIB02 y=b0+b1*AIBllb AIBllb-1
16.544 2.863 0.719 4.364 193.589 - -
BA5 LN(y)=b0+b1*LN(DBH)+b2*LN(CL) n/a 0.276 2.903 0.794 -2.718 0.864 0.496 -
AIB06 y=b1*AIBbhb2
*mLCR - 2.906 2.906 0.760 - 65.956 0.495 -
AIB04 y=b1*AIBllbb2
- 2.925 2.925 0.757 - 83.657 0.583 -
AIB13 y=b1*AIBbhb2
*LCRllb - 2.932 2.932 0.756 - 67.094 0.670 -
SA6 y=b1*SAbhb2
SAbh-2
0.028 2.934 0.689 - 0.043* 1.184 -
SA1 y=b0+b1*SAbh SAbh-2
0.028 2.949 0.686 -1.730* 0.119 - -
AIB08 y=b1*AIBbhb2
*mLCRb3
- 2.968 2.968 0.761 - 66.018 0.501 0.941
SA10 LN(y)=b0+b1*LN(SAbh)+b2*LN(CL) n/a 0.285 3.005 0.702 -2.622* 0.118 2.237* -
SA11 y=b0+b1*SAbh+b2*LCRcb SAbh-2
0.028 3.011 0.687 -2.030 0.116 1.495* -
SA2 y=b0+b1*Sallb SAllb-1
0.340 3.051 0.700 0.764* 0.126 - -
SA7 y=b1*SAllbb2
SAllb-1
0.341 3.055 0.699 - 0.183* 0.935 -
AIB09 y=b1*AIBbhb2
*CL AIBbh-1
13.654 3.064 0.701 - 1.923 0.193 -
CL3 y=b1*CLb2
*dbhob - 3.165 3.165 0.715 - 0.130 0.468 -
BA2 y=b1*BAobb2
BA-1
12.507 3.166 0.699 - 71.374 0.683 -
CL2 y=b0+b1*CL+b2*dbhob - 3.168 3.168 0.702 -5.745 0.679 0.352 -
CL1 LN(y)=b0+b1*LN(CL) n/a 0.301 3.168 0.660 -1.725 0.930 - -
SA9 LN(y)=b0+b1*LN(SAllb) n/a 0.301 3.168 0.660 -1.725 0.930 - -
BA1 y=b0+b1*BAob BA-2
50.121 3.213 0.699 2.483 126.014 - -
AIB12 y=b1*AIBbhb2
*LCRcb - 3.252 3.252 0.715 - 68.899 0.399 -
SA8 LN(y)=b0+b1*LN(SAbh) n/a 0.310 3.263 0.649 -3.185 1.185 - -
AIB01 y=b0+b1*AIBbh - 3.558 3.558 0.625 4.293 121.172 - -
AIB03 y=b1*AIBbhb2
- 3.614 3.614 0.630 - 68.926 0.621 -
AIB05 y=b1*AIBbh*mLCR AIBbh-1
16.218 3.639 0.577 - 245.861 - -
AIB07 y=b1*AIBbh*mLCRb2
- 4.066 4.066 0.700 - 194.300 0.488 -
AIB11 y=b1*AIBbh*LCRllb - 4.227 4.227 0.720 - 233.649 - -
AIB10 y=b1*AIBbh*LCRcb - 4.805 4.805 0.704 - 277.818 - -
* Denotes a parameter coefficient not different than 0 at the 95% confidence level.
Note: Models are ranked by Furnival’s Index values for weighted and log models, and by
root mean square error values for unweighted models.

141. 127
Appendix C
STEM-ANALYZED TREE RING BREAST HEIGHT CHRONOLOGIES
LA 63
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 6
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 61 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 71 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1. Breast height ring chronology of 80 northern white-cedar trees from central
and northern Maine stem-analyzed for early stem development patterns. Note the pattern
of constant or increasing ring widths through time in several sample trees. Dashed
vertical lines indicate date of release. Titles provide sample group (LA=leaf area sample,
M=Maibec sample), tree identification number, and gap origin, if present (G=Gap).

142. 128
LA 161 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 165 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 176
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 180
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 186
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 187
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

143. 129
LA 188
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 211 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 215 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 408
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 576
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 578 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

144. 130
LA 589
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 590
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 598
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
LA 600
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 01
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 02
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

145. 131
M 05
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 03
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 04
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 06
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 08
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 12
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

146. 132
M 13
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 16
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 14 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 17
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 18
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

147. 133
M 19
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 20
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 21
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 22
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 23
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 24
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

148. 134
M 26
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 32
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 25
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 28
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 28
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 33
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

149. 135
M 34
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 36
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 37 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 35
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 39
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 40
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

150. 136
M 41
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 42 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 43
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 44
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 45
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 46
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

151. 137
M 47
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 48 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 50
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 49
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 51
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 52 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

152. 138
M 54
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 53
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 55
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 56
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 57
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

153. 139
M 58
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 63
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 59
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 62
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 64
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 66 G
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

154. 140
M 70
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 73
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 67
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
M 72
0
1
2
3
1700 1750 1800 1850 1900 1950 2000 2050
Radialincrement(mm)
Figure C.1 Continued.

155. 141
Appendix D
MEAN ANNUAL INCREMENT STEM PROFILES
Figure D.1. Mean annual area increment as a function of disc height in 25 stem-analyzed
northern white-cedar trees. A dashed horizontal line denotes the lowest live branch. Note
the high variability of area growth and conspicuous basal flare in many sample trees.
Trees 9, 151, 407, 609 had central decay that did not allow for mean area increment to
include early growth increment. Lack of early growth years led to an inflated mean area
growth in the lower portions of the main stem in these trees.

156. 142
Tree 06 Tree 09
Tree 61 Tree 63
Tree 71 Tree 73
Tree 151
5
10
15
5
10
15
5
10
15
5
10
15
6 12 18 6 12 18
Mean Annual Area Increment (cm2
)
DiscHeight(m)
Tree 161
Figure D.1 Continued.

157. 143
Tree 165
Tree 180 Tree 186
Tree 188
Tree 215
6 12 18 6 12 18
Mean Annual Area Increment (cm
2
)
5
10
15
5
10
15
5
10
15
5
10
15
DiscHeight(m)
Tree 176
Tree 187
Tree 211
Figure D.1 Continued.

158. 144
Tree 407 Tree 408
Tree 576 Tree 578
Tree 589 Tree 590
Tree 598 Tree 600
6 12 18 6 12 18
Mean Annual Area Increment (cm2
)
5
10
15
5
10
15
5
10
15
5
10
15
DiscHeight(m)
Figure D.1 Continued.

159. 145
Tree 609
6 12 18
Mean Annual Area Increment (cm
2
)
5
10
15DiscHeight(m)
Figure D.1 Continued.

160. 146
BIOGRAPHY OF THE AUTHOR
Phil was born in suburban Niagara Falls, NY in 1979. Camping, fishing, and
hiking trips with his family led him down the path to pursuing education in natural
resources. Phil attended SUNY Morrisville to receive an A.A.S. degree in Natural
Resources Conservation. He transferred to SUNY College of Environmental Science and
Forestry and received a B.S. degree in Natural Resources Management. During the
summers of his undergraduate education, Phil worked for the New York State
Department of Environmental Conservation Camp program in the Adirondack region.
His time there was spent courting his future wife, instructing forest ecology lessons, and
leading hiking and paddling expeditions for 15- to 17-year-old campers.
Upon graduation from SUNY-ESF, Phil found few employment opportunities for
the type of forestry he had hoped to practice. He became a self-employed consultant
forester in the Catskill Mountains of New York and managed nearly 10,000 acres for
private landowners. A chance meeting with Dr. Chris Nowak during the summer of 2002
brought Phil back to graduate school at SUNY-ESF in January of 2003 to pursue a M.S.
degree studying even-aged, mixed-species silviculture in northwestern Pennsylvania.
Upon completion, Phil moved to Orono, Maine and immersed himself in conifers and
black flies to study ecology and silviculture of northern white-cedar.
Phil currently lives in Old Town, Maine with his wife, Jessica, and his dog,
Bacchus. He is a candidate for the Doctor of Philosophy degree in Forest Resources from
The University of Maine in May, 2008.