ENVIRONMENTAL EFFECTS ON THE GROWTH OF YOUNG LONGLEAF PINE 
 
 
 
 
 
 
 
Except where reference is made to the work of others, the work described in this thesis is 
my own or was done in collaboration with my advisory committee.  This thesis does not 
include proprietary or classified information. 
 
 
 
 
______________________________________ 
John Christopher Gilbert 
 
 
 
 
 
 
Certificate of Approval: 
 
 
_________________________                                          _________________________ 
William D. Boyer                                                               Ralph S. Meldahl, Chair 
Affiliate Professor                                                              Associate Professor (Emeritus)                           
USDA Forest Service                                                         Forestry and Wildlife Sciences           
 
_________________________                                          _________________________ 
John S. Kush                       Dean H. Gjerstad 
Research Associate IV               Professor 
Forestry and Wildlife Sciences              Forestry and Wildlife Sciences 
 
 
_________________________                                          _________________________ 
David M. Carpenter             George T. Flowers  
Associate Professor             Interim Dean 
Mathematics and Statistics            Graduate School 
 
 
ENVIRONMENTAL EFFECTS ON THE GROWTH OF YOUNG LONGLEAF PINE 
 
 
 
 
 
John Christopher Gilbert 
 
 
 
 
A Thesis 
 
 
Submitted to 
 
 
the Graduate Faculty of 
 
 
Auburn University 
 
 
in Partial Fulfillment of the 
 
 
Requirements for the 
 
 
Degree of 
 
 
Master of Science 
 
 
 
 
 
 
Auburn, Alabama 
May 10, 2007 
 
   
 iii
 ENVIRONMENTAL EFFECTS ON THE GROWTH OF YOUNG LONGLEAF PINE 
 
 
 
 
 
John Christopher Gilbert 
 
 
 
 
 
 
Permission is granted to Auburn University to make copies of this thesis at its discretion, 
upon request of individuals or institutions and at their expense.  The author reserves all 
publication rights. 
 
 
 
 
 
 
 
 
__________________________ 
                                                                        Signature of Author                                      
 
 
 
__________________________ 
                                                                        Date of Graduation 
 
 
 
 
 
 
 
 
 
 
   
 iv
VITA 
 
 John Christopher Gilbert, son of Cynthia (Hurd) Gilbert and the late Robert Porch 
Gilbert, was born January 10, 1980, in Sylacauga, AL.  He graduated from Benjamin 
Russell High School in 1998.  He attended Central Alabama Community College in 
Alexander City, AL and graduated in May of 2001 with an Associates of Arts.  He then 
entered Auburn University in May of 2001, and graduated Summa Cum Laude with a 
Bachelor of Science degree in Forestry in May of 2003.  Directly following graduation in 
May of 2003, he entered Graduate School, Auburn University. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
   
 v
THESIS ABSTRACT 
 
ENVIRONMENTAL EFFECTS ON THE GROWTH OF YOUNG LONGLEAF PINE 
 
 
John C. Gilbert 
 
Master of Science, May 10, 2007 
(B.S., Auburn University, 2003) 
(A.A., Central Alabama Community College, 2001) 
 
Typed Pages 116 
 
Directed by Ralph S. Meldahl 
 
 A study to determine the effects of environmental conditions on the growth of 
young longleaf pine (Pinus palustris Mill) was initiated in 1969 on the Escambia 
Experimental Forest near Brewton, Alabama.  Forty young longleaf trees initially ranging 
from 1 to 1.5 meters in height were measured from 1969 through 1981.  The trees were 
evenly divided between two soil types.  From 1969 to 1970, height and diameter 
measurements were recorded once to four times weekly during the growing seasons and 
once a month during the dormant seasons.  Daily height growth measurements were 
recorded in the morning and again in the evening during the peaks of these two growing 
seasons to determine diurnal and nocturnal growth.  To test the effects shading on 
growth, cheesecloth was suspended over ten randomly selected trees from each soil type 
during the first growing season.  Follow up height and diameter measurements were 
recorded periodically from 1971 through 1981.  All environmental conditions were 
recorded by weather station instruments located on the site.
   
 vi
 The shading treatment did not have a significant effect on either height or 
diameter growth.  Soil type did have a significant effect on diameter growth.   Dormant 
season temperature and precipitation explained more of the variability in yearly height 
and diameter growth than any other environmental conditions measured.  Height growth 
during the peak of the growing season accounted for more than 30% of yearly height 
growth.  Temperature was the most influential environmental variable affecting height 
growth during the peak of the growing season.
   
 vii
 
ACKNOWLEDGEMENTS 
 
The author would like to thank his committee including Dr. Ralph Meldahl, Dr. John 
Kush, Dr. William Boyer, Dr. Dean Gjerstad, and Dr. Mark Carpenter for their guidance, 
insight, and patience throughout this research investigation.  He would also like to thank 
the USDA Forest Service for funding the initial project, Dr. William Boyer for initiating 
the project, T.R. Miller for leasing the land for the project, George Ward and the staff of 
the Escambia Experimental Forest for their work in collecting data, Dr. Richard Brinker 
for early funding assistance to complete the project, Andy Zutter, Anshu Shrestha, and 
Arpi Shrestha, student workers of the Longleaf Pine Stand Dynamics Lab, for their 
assistance in entering data.  Finally, he would like to thank family members Krista 
Gilbert and Cynthia Gilbert for their support, patience, and guidance during this research 
effort. 
 
 
 
 
 
 
 
   
 viii
Style manual or Journal used:  Canadian Journal of Forest Research 
 
 
 
Computer software used:  Microsoft Word? and Excel? 2002, SAS System release? 
9.1, ESRI ArcGIS? 8.3 
   
 ix
 
TABLE OF CONTENTS 
 
LIST OF TABLES............................................................................................................. xi 
 
LIST OF FIGURES ......................................................................................................... xiv 
 
1.  INTRODUCTION ......................................................................................................... 1 
 
2.  OBJECTIVES................................................................................................................ 4 
 
3.  LITERATURE REVIEW .............................................................................................. 5 
3.1  Growth and the Environment................................................................................. 5 
3.2  Flushes and Seasonal Height Growth Patterns .................................................... 11 
 
4.  METHODS .................................................................................................................. 14 
4.1  Location and Layout ............................................................................................ 14 
4.2  Shading Treatment............................................................................................... 16 
4.3  Growth Measurements......................................................................................... 17 
4.4  Environmental Measurements ............................................................................. 18 
4.5  Statistical Procedures........................................................................................... 20 
4.6  Weather and Environmental Conditions.............................................................. 23 
 
5.  RESULTS AND DISCUSSION.................................................................................. 26 
5.1  1969 through 1970 Height Growth...................................................................... 26 
5.2  Flushes ................................................................................................................. 33 
5.3  Needle Lengths .................................................................................................... 38 
5.4  1969 through 1970 Diameter Growth.................................................................. 41 
5.5  Growing Seasons ................................................................................................. 48 
5.6  Diurnal and Nocturnal Growth ............................................................................ 51 
5.7  Height and Diameter Yearly Measurements........................................................ 58 
 
6. CONCLUSIONS........................................................................................................... 65 
 
LITERATURE CITED ..................................................................................................... 69 
 
APPENDICES .................................................................................................................. 75 
Appendix 4.1.1 Soil core information from neutron probe tubes............................... 76 
Appendix 4.4.1 Soil moisture plots and precipitation correlation output................... 77 
Appendix 4.6.1  Precipitation and temperature over the range of available data....... 81 
Appendix 5.1.1 Height growth analysis and plots...................................................... 82
   
 x
Appendix 5.4.1 Diameter growth plots....................................................................... 91 
Appendix 5.6.1 Diurnal and nocturnal growth example regression ........................... 94 
Appendix 5.7.1  Yearly measurements analysis and plots ......................................... 98 
 
 
 
 
   
 xi
LIST OF TABLES 
 
Table                                     Page 
     
Table 4.6.1    Total precipitation and mean temperature by year from                                                    
1969 through 1981 ...................................................................................... 24 
 
Table 4.6.2    Total precipitation and mean temperature by month from                                                  
1969 through 1970 ...................................................................................... 25 
 
Table 5.1.1    Averge height growth during 1969 and 1970 intervals .............................. 27 
 
Table 5.1.2    Average height growth during 1969 and 1970 interval from                                         
March 1969 to December 1970................................................................... 28 
 
Table 5.1.3    Testing overall height growth over monthly interval from                                 
March 1969 to December 1970 using GLM procedure.............................. 28 
 
Table 5.1.4    Testing 1969 monthly height growth from March 1969                                         
to December 1970 using repeated measures procedure.............................. 29 
 
Table 5.1.5    Testing overall height growth over biweekly interval                                                     
from March 1969 to December 1970 using GLM procedure ..................... 29 
 
Table 5.1.6    Testing biweekly height growth over biweekly interval                                                         
from March 1969 to December 1970 using repeated                                   
measures procedure..................................................................................... 30 
 
Table 5.1.7    Testing overall height growth over 1969 weekly interval 
                      from March to October using GLM procedure........................................... 30
 
Table 5.1.8    Testing height growth over 1969 weekly interval 
                      from March to October using repeated measures procedure ...................... 31
 
Table 5.1.9    Testing overall height growth over 1970 weekly interval 
                      from March to November using GLM procedure....................................... 31
 
   
 xii
 
Table 5.1.10  Testing weekly height growth over 1970 weekly interval 
                      from March to November using repeated measures procedure .................. 32
 
Table 5.2.1    Bud and flush regressions of the sampled trees for 1969 and 1970 ........... 37
 
Table 5.3.1    Average needle growth of the sampled trees during 1970.......................... 39 
 
Table 5.4.1    Average diameter growth during 1969 monthly 
                      from March 1969 to December 1969.......................................................... 41 
 
Table 5.4.2    Testing overall diameter growth over monthly interval 
                      from March 1969 to December 1969 using GLM procedure ..................... 42 
 
Table 5.4.3    Average diameter growth during 1970 monthly interval 
                      from January 1970 to December 1970........................................................ 42 
 
Table 5.4.4    Testing overall diameter growth over monthly interval                                           
from January 1970 to December 1970 using GLM procedure ................... 43 
 
Table 5.4.5    Testing monthly diameter growth from March 1969 to 
                      December 1970 using repeated measures procedure.................................. 43 
 
Table 5.4.6    Testing monthly diameter growth from March 1969                                                  
to December 1970 without tree 27 using repeated measures procedure..... 44 
 
Table 5.4.7    Testing 1969 diameter growth over the monthly interval                                      
from March to December without tree 27 using repeated                                
measures procedure..................................................................................... 44 
 
Table 5.4.8    Testing 1970 diameter growth over the monthly interval                                             
from March to December without tree 27 using repeated                                
measures procedure..................................................................................... 45 
 
Table 5.4.9    Testing diameter growth over biweekly interval from                                              
March 1969 to December 1969 using repeated measures procedure ......... 45 
 
Table 5.4.10  Testing diameter growth over biweekly interval from                                                      
March 1970 to December 1970 using repeated measures procedure ......... 46 
 
Table 5.4.11  Testing diameter growth over 1969 weekly interval from March to 
                      October using repeated measures procedure............................................... 46 
 
Table 5.4.12  Testing weekly diameter growth over 1970 weekly interval                             
from March to November using repeated measures procedure .................. 47 
 
   
 xiii
Table 5.5.1    Percent of average annual height and diameter growth  by                                                
month from 1969 through 1970 .................................................................. 49 
 
Table 5.6.1    Diurnal and nocturnal growth regression keywords................................... 54 
 
Table 5.6.2    1969 mean diurnal growth rate (HDR) regressions.................................... 55 
 
Table 5.6.3    1969 mean nocturnal growth rate (HNR) regressions ................................ 55 
 
Table 5.6.4    1969 mean 24 hour growth rate (H24R) regressions.................................. 56 
 
Table 5.6.5    1970 mean diurnal growth rate (HDR) regressions.................................... 56 
 
Table 5.6.6    1970 mean nocturnal growth rate (HNR) regressions ................................ 57 
 
Table 5.6.7    1970 mean 24-hour growth rate (H24R) regressions.................................. 58 
 
Table 5.7.1    Testing overall height growth from 1969 through 1981                                            
using GLM procedure................................................................................. 60 
 
Table 5.7.2    Testing overall diameter growth from 1972 through 1981                                           
using GLM procedure................................................................................. 61 
 
Table 5.7.3    Testing overall diameter growth from 1972 through 1981                                               
using nonparametric Kruskal-Wallis Test .................................................. 61 
 
Table 5.7.4    Yearly remeasurement regression keywords .............................................. 62 
 
Table 5.7.5    Yearly height and diameter regressions...................................................... 63 
  
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
   
 xiv
LIST OF FIGURES 
 
Figure                                     Page 
 
Figure 4.1.1  Study site layout of sampled trees and soil moisture tubes......................... 15 
 
Figure 5.2.1  Average flush lengths of the sampled trees during 1969 ............................ 34 
 
Figure 5.2.2  Average flush lengths of sample trees during 1970 .................................... 35 
 
Figure 5.3.1  Average needle lengths of the sampled trees during 1970.......................... 40 
 
Figure 5.5.1  Percent of average monthly height and diameter growth                                                          
plotted by each month data were collected in 1969 and 1970 .................... 50 
 
Figure 5.7.1  Average diameter measurements by year and soil type 
                      from 1972 through 1981 ............................................................................. 64 
 
 
 
 
 
 
 
 
   
 1
1.  INTRODUCTION 
 Trees, like all terrestrial plants, are dependent on sunlight, carbon dioxide, 
oxygen, water, and nutrients to survive and to grow.  Growth on the most basic scale can 
be defined as an irreversible increase in the size or number of cells (Kramer and Boyer 
1995).  Growth and maintenance are the main purposes of a tree?s metabolism and 
essential to a tree?s survival (Kramer and Boyer 1995).  However, growth is not achieved 
without resistance from environmental stresses.  Changes in the environment result in 
changes in a tree?s internal physiological processes like photosynthesis, respiration, and 
absorption of water and minerals, which in turn affect growth (Kozlowski and Pallardy 
1997b).  The desire to identify and understand the many relationships that exist between 
the environment and tree growth has influenced numerous research efforts, but often the 
answers found lead to more questions.  
 Longleaf pine (Pinus palustris Mill.) is a keystone species across the landscape of 
the southern United States.  It is estimated that the longleaf pine ecosystem once covered 
92 million acres of Southeast, but this vast acreage has been reduced to about 3% of what 
it once was (Frost 1993).  Longleaf pine was removed for its timber value, to establish 
cropland, and to plant species more suited to pulpwood production, like loblolly pine 
(Pinus taeda L.). 
Researchers in the 1930?s began to look at how the growth of longleaf pine was 
related to variation in the environment.  These studies concentrated mostly on 
   
 2
relationships between diameter growth and the environment.  Precipitation and 
temperature data were the only long-term environmental data used, and tree cores were 
examined to determine patterns of past growth (Lodewick 1930).  Diameter growth was 
found to be positively correlated with precipitation and negatively related to temperature 
(Coile 1936).  Lodewick (1930) also found positive correlations between diameter growth 
and precipitation.  Results from these studies raised even more questions about what 
other environmental variables were affecting tree growth because temperature and 
precipitation did not explain all the variance in diameter growth.  Since the majority of 
the research at this time had been conducted with tree cores, evaluating relationships 
between height growth and the environment would require more field measurements over 
time.   
 Boyer (1970) reported on the relationships between shoot growth patterns of 
young loblolly pine and the environment.  This research showed that the most important 
environmental variable affecting height growth was degree-hour heat accumulations 
above threshold temperature.  In this context, threshold temperature is the temperature 
that must be reached for growth to occur.  Solar radiation was also an important variable 
associated with diurnal growth (Boyer 1970).   Boyer (1976), working on the same study, 
found that threshold temperature and growth rate were the most influential variables 
affecting seasonal height growth. 
 In 1969, Boyer initiated a similar study in a young longleaf pine stand in 
southern Alabama.  The main objective of this effort was to try to understand 
relationships between growth of young longleaf pine and the surrounding environment.  
The data were recorded at the Escambia Experimental Forest (EEF) near Brewton, 
   
 3
Alabama on a middle Coastal Plain site from 1969 through 1981.  All of the data were 
collected, but due to a lack of time and funding the study was not completed.  In 
cooperation with Dr. Boyer, this data set was acquired in May of 2003, and the analysis 
was started.  The advances in computer technology since the initiation of this study were 
very beneficial in the completion of this research effort.   
 
 
 
 
   
 4
2.  OBJECTIVES 
 The purpose of this study is to determine the relationships between height and 
diameter growth of young longleaf pine and natural variations in selected environmental 
conditions.  The following questions will be addressed: 
1. What does an examination of seasonal height and diameter growth reveal about 
young longleaf pine growth patterns? 
2. What impact did shading and soil type have on young longleaf pine height and 
diameter growth? What relationships existed between the selected environmental 
conditions and young longleaf pine growth? 
3. Which environmental condition had the greatest effect on young longleaf pine 
growth? 
4. What interactions occurred among the selected environmental conditions at the 
site level, and how did these interactions affect the height and diameter growth of 
young longleaf pine? 
   
 5
3.  LITERATURE REVIEW 
3.1  Growth and the Environment 
 Kozlowski and Pallardy (1997b) cite Klebs, a German plant physiologist, as the 
first person to speculate that plant growth is affected by heredity and the environment.  
Changes in the environment result in changes of internal processes and conditions in 
plants, which in turn affects growth (Kozlowski and Pallardy 1997b).  However, instead 
of a cause and effect relationship between tree growth and the environment, there is a 
complex indirect interaction between the two (Kozlowski and Pallardy 1997b).   
Environmental stresses can be defined as a combination of both biotic and abiotic factors 
(Kozlowski and Pallardy 1997a).  Examples of abiotic stresses include variations in 
rainfall, temperature, moisture, light intensity, fire, pollution, wind, soil structure and 
fertility, while pathogens, insects, and humans are examples of biotic stresses (Kozlowski 
and Pallardy 1997a, Kozlowski and Pallardy 1997b).  To understand more about how 
trees grow, it is necessary to develop a better understanding of the interactions between 
the environment and tree growth (Kozlowski and Pallardy 1997b).  
A plant?s cells continuously swell and shrink depending on changes in water 
content.  This reversible hydration and dehydration of the cells should not be confused
   
 6 
with growth (Kramer and Boyer 1995).  Height growth or apical growth is the elongation, 
division, and maturation of cells in the apical meristem (Kozlowski and Pallardy 1997b).  
Height growth involves a combination of node and internode elongation of one or 
multiple leaders (Kozlowski and Pallardy 1997b).  During the growing season, height 
growth occurs in waves or flushes of growth and can be classified as fixed and/or free 
growth depending on species and the climate (Kozlowski and Pallardy 1997b).  Diameter 
growth or cambial growth occurs by the division of fusiform initial cells in the vascular 
cambium.  The dividing fusiform initial cells differentiate into both xylem and phloem 
cells which collectively make up diameter growth (Kozlowski and Pallardy 1997b).  
Kozlowski and Ward (1961) state that height growth is inconsistent between species and 
even among trees of the same species, but diameter growth curves tend to vary more year 
to year than height growth curves and are more receptive to changes in the environment. 
Longleaf pine seedlings are the only southern pines that exhibit a short shoot 
characteristic.  Brown (1964) defines the seedling as a short shoot with a bud that 
produces a single needle to many needles at the start of the growing season but does not 
elongate itself.  This is commonly referred to as a grass stage seedling (Wahlenberg 1946, 
Boyer 1990).  The seedling will remain in the grass stage until a sufficient root system 
has developed (Outcalt 2000).  Pessin (1935) found an inversely proportional relationship 
between root growth and height growth while seedlings were still in the grass stage.  
When the root-collar diameter reaches about 2.5 cm, height growth is usually initiated 
(Wahlenberg 1946, Boyer 1990).  Once height growth is initiated, seedlings tend to have 
a rapid phase of height growth where they emerge from their ground layer vegetative 
competitors.  At this point the seedlings have developed into saplings and can grow about 
   
 7 
30.5 to 100 cm a year, which varies with environmental conditions including degrees of 
competition and weather conditions (Wahlenberg 1946).   
A more intense understanding of what and how environmental conditions are 
affecting tree growth has the potential to be used to make more effective silvicultural 
prescriptions (Kozlowski 1955).   Manipulating light availability is very important since 
longleaf pines are considered shade intolerant (Wahlenberg 1946, Boyer 1990).  McGuire 
et al. (2001) found a positive correlation between the amount of light available and 
longleaf pine seedling growth.  Barnett (1989) took a different approach by looking at the 
effect of shading on container grown seedlings in the nursery.  He found significant 
differences between the shading treatments and a full sunlight control.   However, he did 
not see significant differences between the 30% and 50% shading treatments.  Palik et al. 
(1997) studied the effects of sunlight and nitrogen on longleaf pines planted in forest gaps 
where light ranged from 40% to 80% of full sunlight.  They found an upward curvilinear 
relationship between the growth rates of the selected trees and the variations in nitrogen 
and sunlight levels.   
 Of the natural resources a tree needs to survive, water has been 
characterized as the most influential (Zahner 1956, Shoulders and Tiarks 1980).  Zahner 
(1956) states that in order for a tree to grow and maintain itself, it needs water more than 
any other raw material.  Shoulders and Tiarks (1980) state that the natural distribution 
and growth of trees are greatly dominated by water availability.  Their study showed that 
longleaf pine has a lower tolerance to fall and winter rainfall than loblolly and slash pine 
(Pinus elliottii Engelm.).  This study also showed the relationship between subsoil 
moisture and height growth to be a third degree polynomial.  At 7% subsoil moisture the 
   
 8 
curve changes from a positive relationship to a negative one, but at about 18% subsoil 
moisture the relationship became positive again (Shoulders and Tiarks 1980).  They also 
observed a threshold of height growth response to spring and summer rainfall at about 
114 to 144cm.  Rayamajhi (1996) found minimum temperature and rainfall to be the most 
important environmental variables affecting changes in longleaf pine growth.  In a study 
on the Coastal Plain of Georgia, Coile (1936) found a positive correlation between 
longleaf pine growth and February through April rainfall.  Lodewick (1930) explored 
relationships between longleaf pine diameter growth and precipitation in western Florida.  
He found no relationship between early growing season precipitation and diameter 
growth but did see a positive correlation between rainfall and growth later in the growing 
season.  Lodewick (1930) saw a negative relationship between springwood growth and 
increases in rainfall during the growing season, but springwood growth had a positive 
relationship with summerwood growth.  Rai (1995) also found a negative relationship 
between rainfall and longleaf pine growth when studying longleaf pine on the Gulf 
coastal plains of Alabama, Georgia, and Florida for five years.  Coile (1936) found a 
negative correlation between average summer temperature from June through August and 
diameter growth, but Lodewick (1930) found no relationship between diameter growth 
and temperature.  Coile (1936) saw variation between annual diameter growth and 
temperature, but stated that when radial growth was linear, good relationships could be 
found between diameter growth and rainfall and temperature.  Coile (1936) stated that the 
relationship between growth and rainfall was curvilinear, but this was dependent on the 
portion of the growing season that was being analyzed. 
 Bengston et al. (1967) looked at 6-year-old slash pines from 14 geographic 
   
 9 
sources on a study site in Florida and found a proportional relationship between rainfall 
and height growth in slash pine.  Treubig (1960) studied longleaf and loblolly pine in 
southern Louisiana and also saw a relationship between longleaf pine growth and rainfall.  
The study showed a significant relationship between longleaf pine growth and the 
number of weeks with less than 2.54 cm of rainfall but did not see that environmental 
factors explained much of the variation in longleaf or loblolly pine growth.   
Boyer (1970, 1976) measured the height growth of 10 loblolly pine trees during 
the growing season of 1967.  Diurnal and nocturnal growth rate measurements were taken 
twice daily during the peak of the growth period.  This research showed that the most 
important environmental variable affecting height growth was degree-hour accumulations 
above threshold temperature, which was 40? F (4.44? C) during the night and 50? F (10.0? 
C) during the day.  Boyer (1976) states that the threshold temperature and the growth rate 
above the threshold temperature explained 94% of the differences in height growth 
among the sampled trees.  Kramer (1957) also found that temperature was very 
influential to height growth.  He found that the difference between the maximum and 
minimum temperature for a day was more significant in explaining height growth than 
either temperature alone.  
   Solar radiation and vapor pressure deficit also seemed to be important variables 
associated with growth (Boyer 1970).  In this case, there was a direct relationship 
between the intensity of solar radiation and threshold temperature for diurnal growth 
while degree-hour accumulation and vapor pressure deficit were significant in explaining 
diurnal growth.  Addington et al. (2004) looked at how vapor pressure deficit affects 
longleaf pine.  They found a negative relationship between stomatal conductance and 
   
 10
vapor pressure deficit, which helps explain changes in whole-plant hydraulics.   
McLaughlin et al. (2003) found growth rates of yellow-poplar (Liriodendron tulipifera 
L.) to be negatively affected by high values of vapor pressure deficit.  They performed a 
study on 10 mature yellow-poplars that were recently released.  The study took place in 
Tennessee and focused on the relationship between diurnal patterns of growth and 
climatic factors.    
Yeh and Wensel (2000) studied the relationship between soil moisture and growth 
of 6 western coniferous species in northern California.  They showed that 67% of the 
growth variations in the measured pines were explained by variations in temperature and 
precipitation.  Kienholz (1941) looked at hardwoods in Connecticut and found no 
statistically significant relationships between rainfall and height growth curves.  Fritts 
(1960) studied hardwoods including the following:  American beech (Fagus grandifolia 
Ehrh.), white oak ( Quercus alba L. ), red oak ( Quercus rubra L.), and sugar maple 
(Acer saccharum Marsh.).  He found the most influential variable affecting the growth 
was temperature, but labeled maximum temperature, soil moisture, percent sunshine of 
the measurement and previous day, relative humidity, and wind speed as important 
variables when considering how growth is affected by the environment (Fritts 1958, 
1960) 
Most studies concerning longleaf pine height and diameter growth have been 
conducted in controlled environments like greenhouses.  There has been a lack of studies 
that concentrate on the environmental effects of height and diameter growth under field 
conditions or natural conditions (Boyer 1970).  Jose et al. (2003) evaluated the effects of 
light, water, and nitrogen on grass stage longleaf pine seedling growth.  The study, 
   
 11
performed in a greenhouse, concluded that seedling growth seemed to be more influenced 
by water and nitrogen than light.  Ramsey et al. (2003) studied the effects of fertilization 
and herbaceous weed control on longleaf pine seedlings.  They concluded that 
herbaceous weed control combined with fertilization would increase seedling height 
growth and decrease the length of the grass stage.  However, fertilization at establishment 
tended to reduce seedling survival, while herbaceous weed control after the first growing 
season had little effect.   
3.2  Flushes and Seasonal Height Growth Patterns  
In contrast to diameter growth, longleaf height growth is somewhat continuous 
and occurs in flushes (Byram and Doolittle 1950).  Allen and Scarbrough (1969) looked 
at seasonal patterns and flushes in young longleaf pine from 1957 to 1969 on a monthly 
and weekly basis.  They observed the majority of growth between March and October 
which they classified as the growing season.  In examining the patterns of growth, they 
observed continuous growth of flushes.  From about the middle of April until the middle 
of July there were always at least two flushes elongating at the same time.  Tepper (1963) 
also saw growth of two flushes at one time in pitch pine (Pinus rigida Mill.) and shortleaf 
pine (Pinus echinata Mill.). Mudano et al. (1992) and Griffing and Elam (1971) looked at 
loblolly pine in the Duke Forest near Durham, NC and northern Mississippi, respectively 
and reported continuous growth of two to three flushes at one time throughout the 
growing season.  However, contrary to the idea of concurrent growth, Eggler (1961) 
recorded consecutive growth when measuring the 4 southern pines in a study in 
Louisiana and Mississippi.  He reported that after one bud stops, another begins to grow 
   
 12
without a ?rest? period.   
The first flush of loblolly pine contains about 30% of the total year?s growth and 
is longer than any of the other flushes from the growing season (Mudano et al. 1992,  
Kozlowski and Pallardy 1997b).  The first flush of the growing season resembles a 
sigmoid curve with a maximum in April.  Allen and Scarbrough (1969) found that the 
curves of the rest of the flushes were not sigmoid in shape for longleaf pine.  In loblolly 
pine, budbreak occurred during mid to late March, 90% of total shoot growth was 
completed by the middle of August, and growth ceased in early November (Mudano et al. 
1992).  They found that the third flush tends to start about the time the first flush stops 
growth while the second flush starts growing as the first flush reaches its peak of growth.   
 The number of flushes during a growing season is correlated with tree vigor 
(Allen and Scarbrough 1969).  This concept leads to the idea that vigor is dependent on 
the environment which implies that the number and duration of flushes is dependent on 
changes in the environment.  The study showed evidence that total height was not totally 
dependent on the number of flushes.  However, they did find that that about 47% of the 
variation in the spring flushes and about 74% of the variation in the summer flushes was 
accounted for by the number of nodes (Allen and Scarbrough 1970).  Bengston et al. 
(1967) observed significant correlations between the length of the bud set prior to the 
growing season and total height growth (r=0.91) and early height growth (r=0.97) in slash 
pine.   Boyer (1970) found a significant regression where length of the bud set prior to the 
growing season explained 78% of the variability in length of the first internode for 
loblolly saplings.  He also found that length of the bud set prior to the growing season 
and the length of the first internode explained 79% and 87%, respectively, of the 
   
 13
variability in the bud set at the end of the growing season for loblolly pine.   Allen and 
Scarbrough (1970) also found that there was a significant correlation between successive 
flushes which they attribute to the similarities in internode elongation in longleaf pine.  
However, Griffing and Elam (1971) did not find significant correlations between flushes 
of the same year or with the previous year. Allen and Scarbrough (1970) reported a 
negative correlation between the spring growth and summer growth of the next year.  
Through experimental treatments, Allen and Scarbrough (1969) found that removing 
foliage did reduce total height growth by half but did not have an effect on the number of 
flushes.  
 
 
 
 
   
4.  METHODS 
4.1  Location and Layout 
This project was installed on the Escambia Experimental Forest (EEF) near 
Brewton, AL in 1969.  The EEF is a 3000-acre forest that was established in 1947, when 
T.R. Miller Mill Company leased the land to the USDA Forest Service for 99 years at no 
cost (Boyer et al. 1997).  The EEF has been used for extensive longleaf pine research 
which has resulted in numerous publications (Boyer et al. 1997). 
The study sample consisted of young naturally regenerated longleaf pines, which 
were a product of the 1955 seed crop.  Forty longleaf pine trees ranging from 1 to 1.5 m  
in height were selected for the study.  Twenty were selected on each of the 2 distinct soil 
types that were present in the stand.  Figure 4.1.1 shows stem map layout of the study 
site.  
The two soil types that separated the trees were Lucy loamy sand and Wagram 
loamy sand.  The taxonomic class for both soils is:  loamy, kaolinitic, thermic Arenic 
Kandiudults (Soil Survey Division, Natural Resources Conservation Service 2003). The 
trees on the Lucy loamy sand were located on the crest of a ridge, and the trees on the 
Wagram loamy sand were located on a slope at the base of the ridge. The soils are very 
similar with an average site index of 20.4 m.  The depth of the A-horizon for the Lucy 
soil can vary from 55.88 cm to 101.6 cm (Mattox et al. 1975).   
14
 
   
Figure 4.1.1  Study site layout of sampled trees and soil moisture tubes  
 
 
 
 
 15
   
 16
For the Wagram soil, the depth of the A-horizon can vary from 50.8 to 68.58 cm (Mattox 
et al. 1975).  These estimates were developed from a representative sample and will vary 
depending on the site.  Soil samples were taken for this study in coordination with the 
collection of soil moisture at each neutron probe tube.   Appendix 4.1.1 is a table of A-
horizon and B-horizon depths for each tube.  The average depth of the A-horizon on the 
site was 121 cm for Lucy soil and 107 cm for Wagram soil.  The average depth to the 
base of the B-horizon was 207 cm for Lucy soil and 226 cm for Wagram soil.  A deeper 
average A-horizon should make the Lucy soil drier than the Wagram soil.  However, the 
sandy clay loam in the shallower B-horizon of the Lucy soil could provide more available 
moisture for roots than the deeper B-horizon of the Wagram loamy sands (Mattox et al. 
1975) 
4.2  Shading Treatment 
 To create artificial shading, half of the trees in each of the two soil types were 
randomly selected for artificial shading.  These trees were shaded with cheesecloth for six 
months during the growing season of the first year of the study.  The shading treatment 
was installed March 28, 1969 and removed September 24, 1969.  The cheesecloth was 
stretched across a one-meter-square frame that was structured to keep the cheesecloth at 
least one meter above the growing tip of the tree.  The structures were periodically 
checked, adjusted, and maintained at one meter above the growing tip.  The cheesecloth 
prevented the growing tips of the shaded trees from receiving direct sunlight during the 
peak of the diurnal cycle.   The mean percent difference in solar radiation between non-
shaded and shaded was 38.5%.   
   
 17
 
4.3  Growth Measurements 
 Initial heights were measured from the ground line to the base of the bud on the 
terminal growing shoot of each tree on January 27, 1969.  The length of the terminal bud 
was also measured.  Separate records were maintained for each new leader. Terminal 
shoots were measured from 2 to 4 times weekly from March to October of 1969 and 
1970.  During the dormant season of both years, heights were measured every two weeks 
or at least once a month.    
 Monthly growth intervals covered 22 months from March 1969 through 
December 1970.  Each monthly interval was 28 days +/- 1 day.  The biweekly intervals 
included 22 measurements in 1969 ranging from March until December and 24 
measurements in 1970 ranging from January to December.  Biweekly growth intervals 
were 14 days +/- 1 day in length.  Weekly intervals included 34 measurements in 1969 
ranging from March through October and 38 measurements in 1970 ranging from March 
through November.  The weekly intervals were 7 days +/- 1 day in length.  From April 28 
to October 28 of 1970 needle lengths were recorded for each tree in addition to height 
measurements.   
 More intensive measurements were taken during April of both years, which was 
the peak of the growing season.  The trees were measured on 19 days during April, 1969 
and 16 days during April, 1970.   Measurements were taken in 1969 from April 9 through 
April 28 and in 1970 from April 6 through April 28.  During April of 1970, 
measurements were not taken on Sunday including April 12
th
, 19
th
, and 26
th
.  Each day, 
   
 18
leaders were measured in the morning from 8:00 to 10:00 a.m. and again in the late 
afternoon from 4:00 to 6:00 p.m.  Height growth occurring between the morning and 
afternoon measurements was labeled as diurnal growth, and height growth occurring 
between the afternoon measurement and the following morning measurement was labeled 
as nocturnal growth (Boyer 1970).  Height growth occurring from the morning to 
morning was labeled as total growth (Boyer 1970).  
 Diameters of each tree were measured in centimeters at 10 cm above the ground 
line with the use of dendrometer bands. The 10 cm height was set because the trees had 
not reached dbh (diameter at breast height) of 1.37 m.  Diameter measurements were 
taken weekly from March thru October of 1969 and 1970.  From the end of October until 
March, measurements were taken every two weeks to a month during 1969 and 1970.  
These measurements can again be divided into monthly, biweekly, and weekly growth 
intervals using the same interval lengths and durations as outlined for height growth.  
 The sample trees were remeasured yearly during the dormant seasons from 1971 
through 1981.  Heights and diameters were recorded for each tree.  All height 
measurements were recorded to the base of the bud on the terminal shoot.  During the 
remeasurement period, diameter measurements were recorded at dbh.  
4.4  Environmental Measurements 
 Environmental conditions recorded included: maximum and minimum air 
temperature, precipitation, soil moisture, relative humidity, wind speed, and solar 
radiation.  An onsite weather station equipped with maximum and minimum 
thermometers and a rain gauge recorded maximum and minimum air temperatures and 
   
 19
precipitation on a daily basis.  Air temperatures were recorded in degrees Fahrenheit and 
converted to degrees Celsius.  A temperature difference can be calculated from the 
recorded maximum and minimum recorded temperatures for each day.  Precipitation data 
were recorded in inches and tenths and were to centimeters.  Data from the weather 
station were available on a daily basis from 1964 through 1990.    
 Soil moisture profiles were also created by using a neutron probe at eight 
locations with depths of 10, 25, 50, 75, 100, 125, 150, and 175 cm.  Four tubes for the 
neutron probe were installed in each soil type, and the neutron probe was recalibrated at 
each soil type.  Neutron probe measurements were taken at weekly intervals from March 
to October of 1969 and 1970.  From the end of October until March, measurements were 
taken once every two weeks to a month.  Eight soil samples were also taken near the 
access tubes used for the neutron probes. The samples were taken as 5.08 cm cores and 
were a depth of 1.9 m and 1 to 1.5 m from the access tubes.  The cores were analyzed at 
the Longleaf Silviculture Laboratory in Brewton, AL.  The data obtained from these 
cores included:  moisture retention capacity, texture, organic matter content, and bulk 
density. 
 The soil moisture calibration data were not available for this study.  To calculate 
percent soil moisture from the neutron counts, a factory curve for the neutron probe that 
best fit the data was used.  Points from the curve were interpolated and used to find the 
equation for the line.  The actual percent soil moisture values will be misleading, but the 
change between percent soil moisture values will be more suitable for this analysis and 
give relative good values.  A correlation analysis was computed between the percent soil 
moisture values and rainfall to see how closely the two sets of data were associated.  The 
   
 20
correlation matrix can be found in Appendix 4.4.1.  The estimated soil moisture values 
were significantly correlated with lagged precipitation.  For analyses used in this study, 
lagged precipitation will be used instead of the estimated soil moisture values. 
 During April of 1969 and 1970, more intensive environmental measurements 
were taken.  These measurements included relative humidity, wind speed, and solar 
radiation measurements.  Environmental variables during the month of April of both 
1969 and 1970 were taken on days corresponding to the dates that heights were recorded.  
  A hygrothermograph was used to measure air temperature in degrees Fahrenheit 
and humidity.  Mean temperatures could be calculated for each diurnal, nocturnal, and 
24-hour period.  Relative humidity was calculated in percent for each period by summing 
values for each hour and dividing by time (Boyer 1970).  Degree-hour heat sums or 
degree-hour accumulations were calculated by summing degrees above threshold 
temperatures of 40?, 45?, 50?, and 55? F for each period (Boyer 1970).  Vapor pressure 
deficits were calculated in units of millimeters of mercury from mean temperature and 
relative humidity.  An anemometer was positioned at 2.25 meters from the ground to 
record wind speeds.  Winds speeds were recorded in total miles for each diurnal and 
nocturnal period.  A pyrheliograph was used to measure solar radiation in both shaded 
and unshaded areas.  Solar radiation was measured in langleys (cal/cm
3
). 
4.5  Statistical Procedures 
 All statistical procedures were executed in SAS (Statistical Analysis System) 
software version 9.1 (SAS 2003).   All analyses were conducted at the 0.05 level of 
significance.   Analyses of variances were conducted to test for the effects of shade 
   
 21
treatments and soil type conditions over various growth rates using PROC GLM and 
PROC MIXED in SAS (SAS 2003).  PROC MIXED was used to conduct a repeated 
measures analysis using tree as the random factor (SAS 2003).  The repeated measures 
analyses were conducted with the measurements from each interval and with consecutive 
growth from the initial measurement.  PROC MIXED was chosen because it allows 
missing data and an unbalanced data structure (Littell et al. 1996).  All date variables 
used in the repeated measures analysis were converted to Julian days using SAS.  SAS 
converts the date into the number of days since January 1, 1960 (Cody and Smith 1997).  
Total growth over each interval was also tested using a PROC GLM (SAS 2003).  Type 
III sums of squares were presented for ANOVA effects because unbalanced data were 
used (SAS 2004).  The results of the different tests were compared before final 
conclusions were determined.   Significant interactions were tested further using least 
significant difference (LSD) multiple comparison procedures and within group t-tests 
(Ramsey and Schafer 2002, SAS 2003). 
  Regression and correlation procedures were used to determine the relationships 
between recorded growth and environmental conditions.  Height and diameter growth 
were the dependent values and the environmental conditions were the independent 
variables.  Correlation procedures were used to evaluate relationships between the 
independent variables.  Regressions between growth and each explanatory variable were 
evaluated individually before attempting selection searches.  Both forward and backward 
step-wise regression procedures were used, resulting in significant simple and multiple 
regressions (Neter et al. 1996).  Simple linear and multiple regressions were evaluated 
using the coefficient of multiple determination (R
2
) (Neter et al. 1996).  Multiple 
   
 22
regressions were also evaluated using the adjusted coefficient of multiple determination 
(adjusted- R
2
) because it is not inflated by additional variables in the model (Neter et al. 
1996).     
 All statistical tests were evaluated to confirm any potential violations from 
assumptions.  To evaluate normality of residuals, a normal probability plot or a normal 
QQ plot was used from PROC UNIVARIATE (Neter et al. 1996, Ramsey and Schafer 
2002, SAS 2004).  A normal QQ plot is a plot of the ordered residuals of a selected 
variable against a theoretical normal distribution, where the points should lie on a straight 
line (SAS 2004).  Deviations from normality occur when points do not lie on the line.  
The distribution can be considered long-tailed when there are large gaps between 
residuals in the tails where points are off the straight line on both sides (Ramsey and 
Schafer 2002).  The distribution is considered skewed where there are gaps between 
residuals on one side of the line and shorter on the other side (Ramsey and Schafer 2002).  
Outliers can also be detected in the QQ plot where one or a few points are far off the 
straight line and the rest follow the normal distribution (Ramsey and Schafer 2002).  The 
Anderson-Darling test from PROC UNIVARIATE was also used to determine normality 
of the residuals (SAS 2004).  Results from the QQ plot and from PROC UNIVARIATE 
were compared to determine if normality was suspect.  If the assumption of normality 
was violated, data transformations were applied.  If suitable transformations did not 
correct the violation of normality, the NPAR1WAY procedure in SAS was used to 
calculate the nonparametric Wilcoxon Rank Sum Test and the Kruskal-Wallis Test, 
where p-values less than the 0.05 level of significance were considered significant (Cody 
and Smith 1997, Ramsey and Schafer 2002).   
   
 23
 Both Levene?s and Bartlett?s tests were used to determine homogeneity of 
variance (Ramsey and Schafer 2002, SAS 2004).   A p-value less than the 0.05 level of 
significance shows deviation from the homogeneity of variance (Ramsey and Schafer 
2002).  Plotting the residuals against the predicted values is also useful in determining 
homogeneity of variance, outliers, and if linear models are appropriate for regressions 
(Neter et al. 1996, Ramsey and Schafer 2002).  Plots should show residuals evenly 
spaced above and below the zero line across the values without systematic deviations.  
Systematic deviations include increases in spread as one variable increases.  To correct 
outliers and other deviations in the homogeneity of variance, transformations are used 
(Neter et al. 1996, Ramsey and Schafer 2002).  
4.6  Weather and Environmental Conditions 
 Onsite weather station data were available from 1964 through 1990.  Table 4.6.1 
includes data on total yearly precipitation, average maximum temperature, and average 
minimum temperature for each year during the actual years of the study.  From 1969 
through 1981, the average precipitation for the area was 160 cm.  The driest year was 
1971 with only 111.53 cm of precipitation, and 1975 was the wettest year with a total of 
229.74 cm.   The average maximum temperature during this study was 26
?
C, while the 
average minimum temperature was 12
?
C.  The highest average maximum temperature 
was in 1978, while 1976 had the lowest average minimum temperature. A table providing 
the total yearly precipitation, average maximum temperature, and average minimum 
temperature for the entire range of the data, 1964 through 1990 can be found in the 
Appendix 4.6.1. 
 
   
 24
Table 4.6.1  Total precipitation and mean temperature by year from 1969 through 1981 
 
Year 
Precipitation 
(cm) 
Avg. Maximum 
Temperature (?C)
Avg. Minimum 
Temperature (?C)
1969 176.78 25 11 
1970 195.83 25 11 
1971 111.53 25 12 
1972 145.92 27 13 
1973 155.96 26 13 
1974 136.14 26 12 
1975 229.74 24 11 
1976 124.08 25 11 
1977 140.84 27 13 
1978 144.53 29 15 
1979 172.34 25 12 
1980 167.89 26 12 
1981 126.37 26 11 
 
  
 Table 4.6.2 provides total monthly precipitation, average maximum temperature, 
and average minimum temperature for each month during 1969 and 1970.  Peaks in 
precipitation occurred during July of 1969 and during June of 1970.  Average maximum 
temperatures were at the highest in June of 1969 and July of 1970.  December was the 
coldest month in 1969, while, February was the coldest in 1970. 
 
 
 
 
 
 
   
 25
Table 4.6.2  Total precipitation and mean temperature by month from 1969 through 1970 
 
 
Year Month 
Precipitation 
(cm) 
Avg. Maximum 
Temperature (?C) 
Avg. Minimum 
Temperature (?C) 
1969 1 5.46        16 4 
1969 2 10.16        18 4 
1969 3 21.46        18 4 
1969 4 13.21        26 12 
1969 5 16.13        29 15 
1969 6 5.46        34 19 
1969 7 33.02        33 21 
1969 8 26.16        31 19 
1969 9 8.13        30 16 
1969 10 7.37        26 13 
1969 11 4.45        20 3 
1969 12 25.78        16 2 
1970 1 8.64        13 1 
1970 2 12.57        17 0 
1970 3 25.53        20 6 
1970 4 2.03        28 13 
1970 5 23.75        30 14 
1970 6 34.29        30 18 
1970 7 11.81        33 20 
1970 8 30.99        32 21 
1970 9 3.94        33 19 
1970 10 22.61        25 13 
1970 11 5.33        19 3 
1970 12 14.35        19 5 
 
 
 
 
 
 
 
 26
5.  RESULTS AND DISCUSSION 
5.1  1969 through 1970 Height Growth 
Evaluating differences in height growth over the different intervals is necessary to 
see if the recorded environmental conditions and treatments are affecting height growth.  
Each interval (monthly, biweekly, and weekly) covers different portions of the growing 
seasons.  Patterns of growth over the growing seasons show that trees do not grow at the 
same rates during the entire growing season.  Growth rate tends to vary over the growing 
season which could raise questions about how relationships between growth and the 
environment change over the same period.  Table 5.1.1 displays mean height growth for 
1969 and 1970 individually, and Table 5.1.2 displays mean height growth over the two 
year measurement period from March of 1969 through December 1970. 
Overall height growth was first tested over the entire monthly interval.  Table 
5.1.3 shows the effects table from this test.  There were no significant differences 
between shade treatments or soil conditions, but there was a significant interaction 
between the soil and shade variables.  Within group comparisons were then used to 
evaluate the significant interaction.  The normality assumption was checked for each test 
using a QQ plot and the Anderson-Darling test generated from PROC UNIVARIATE 
(SAS 2003, SAS 2004).  Barlett?s and Levene?s tests were used to check for violations of 
constant variance.   Unless otherwise stated, the assumptions were not violated. See 
Appendix 5.1.1 for a detailed description of the statistical tests and results.
 
 
 
 27
Table 5.1.1  Average height growth during 1969 and 1970 intervals  
 
Year   Soil Type Average Height Growth (m) 
1969  Lucy 1.01 a 
  Wagram 0.99 a 
    
1970  Lucy  1.09 b 
  Wagram 1.12 b 
    
Year   Shade Treatment Average Height Growth (m) 
1969  Not Shaded 0.99 c 
  Shaded 1.01 c 
    
1970  Not Shaded 1.09 d 
  Shaded 1.12 d 
    
Year   Shade/Soil Combination Average Height Growth (m) 
1969  NW (Not Shaded, Wagram) 1.02 e 
  SL (Shaded, Lucy) 1.06 e 
  SW (Shaded, Wagram) 0.96 e 
  NL (Not Shaded, Lucy) 0.97 e 
    
1970  NW (Not Shaded, Wagram) 1.20 f   
  SL (Shaded, Lucy) 1.20 f 
  SW (Shaded, Wagram) 1.04 fg 
    NL (Not Shaded, Lucy) 0.98 g 
Means with the same letter are not significantly different (?=0.05) 
NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 28
Table 5.1.2 Average height growth during 1969 and 1970 interval  
from March 1969 to December 1970  
 
 
Treatment/Condition Average Height Growth (m) 
No Shade 2.10 a 
Shade 2.09 a 
Lucy 2.10 b 
Wagram  2.09 b 
Shade/Soil Combination Average Height Growth (m) 
NW 2.21 c 
SL 2.21 c   
NL 1.99 cd 
SW 1.95 d 
Means with the same letter are not significantly different (?=0.05) 
NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
Table 5.1.3  Testing overall height growth over monthly interval from March 1969 to 
December 1970 using GLM procedure  
 
Source Type III SS Mean Square F Value Pr > F 
SOIL 38.07 38.07 0.05 0.8164 
SHADE 73.33 73.33 0.11 0.7474 
SOIL*SHADE 5448.88 5448.88 7.83 0.0084* 
*Means significantly different (?=0.05) 
Monthly height growth was explored for potential significant differences using 
the repeated measurements design.  Below is the effects table (Table 5.1.4) from the 
PROC MIXED output.  Only time related variables and the shade/soil interaction were 
significantly different.   
 
 
 
 
 
 
 
 
 
 
 29
Table 5.1.4  Testing 1969 monthly height growth from March 1969 to December 1970 
using repeated measures procedure  
 
Effect         DF F Value Pr > F 
TIME 1069 4755.63 <.0001* 
SOIL 1069 0.24 0.6255 
SHADE 1069 0.10 0.7558 
SOIL*SHADE 1069 12.67 0.0004* 
TIME*SHADE 1069 0.10 0.7493 
TIME*SOIL 1069 0.01 0.9277 
TIME*SOIL*SHADE 1069 16.37 <.0001* 
*Means significantly different (?=0.05) 
Testing the biweekly interval resulted in no significant differences between shade 
treatments or soil condition, but the shade/soil interaction was significant (Table 5.1.5).  
Table 5.1.6 contains the effects table of the repeated measures test looking at biweekly 
intervals. Again, only time related variables and the shade/soil interaction were 
significantly different.   
 
Table 5.1.5  Testing overall height growth over biweekly interval from March 1969 to 
December 1970 using GLM procedure  
 
Source DF Type III SS 
Mean 
Square F Value Pr > F 
SOIL 1 199.04 199.04 0.61 0.4408 
SHADE 1.03 1.03 0.00 0.9555 
SOIL*SHADE 1 2746.06 2746.06 8.39 0.0065* 
*Means significantly different (?=0.05) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 30
Table 5.1.6  Testing biweekly height growth over biweekly interval from March 1969 to 
December 1970 using repeated measures procedure  
 
Effect Num DF Den DF F Value Pr > F 
TIME 1 1802 8410.46 <.0001* 
SOIL 1802 0.21 0.6470 
SHADE 1 1802 0.03 0.8595 
SOIL*SHADE 1802 19.88 <.0001* 
TIME*SHADE 1 1802 0.03 0.8576 
TIME*SOIL 1802 0.07 0.7857 
TIME*SOIL*SHADE 1 1802 25.33 <.0001* 
*Means significantly different (?=0.05) 
Weekly intervals were tested for both 1969 and 1970 because the measurements 
were not consecutive.  Tables 5.1.7 and 5.1.8 contain the output for the GLM and 
MIXED procedures, respectively. Only the time variable was significantly different.  This 
shows that the soil shade interaction occurred in the 1970 growth. 
 
Table 5.1.7  Testing overall height growth over 1969 weekly interval  
from March to October using GLM procedure 
  
Source DF Type III SS 
Mean 
Square F Value Pr > F 
SOIL 1 178.99 178.99 0.85 0.3634 
SHADE 9.24 9.24 0.04 0.8354 
SOIL*SHADE 1 176.26 176.26 0.84 0.3670 
*Means significantly different (?=0.05) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 31
Table 5.1.8  Testing height growth over 1969 weekly interval 
 from March to October using repeated measures procedure  
 
 
 
Effect 
Num 
DF Den DF F Value Pr > F 
TIME 1 1383 3692.33 <.0001* 
SOIL 1383 0.77 0.3807 
SHADE 1 1383 0.02 0.9020 
SOIL*SHADE 1 1383 0.84 0.3608 
TIME*SHADE 1 1383 0.01 0.9088 
TIME*SOIL 1 1383 1.39 0.2382 
TIME*SOIL*SHADE 1 1383 1.17 0.2799 
*Means significantly different (?=0.05) 
 Tables 5.1.9 and 5.1.10 contain the pertinent output for the GLM and MIXED 
procedures, respectively, for the 1970 weekly height intervals. No significant differences 
were observed except the shade/soil interaction.  Table 5.1.8 shows that only time and the 
shade/soil interaction were significantly different.   
 
Table 5.1.9  Testing overall height growth over 1970 weekly interval 
 from March to November using GLM procedure  
 
Source DF Type III SS 
Mean 
Square F Value Pr > F 
SOIL 1 204.16 204.16 0.69 0.4136 
SHADE 1 5.76 5.76 0.02 0.8902 
SOIL*SHADE 1 2572.46 2572.46 8.63 0.0059* 
*Means significantly different (?=0.05) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 32
 
 
Table 5.1.10  Testing weekly height growth over 1970 weekly interval  
from March to November using repeated measures procedure  
 
Type 3 Tests of Fixed Effects 
Effect Num  DF Den DF F Value Pr > F 
TIME 1 1437 2399.64 <.0001* 
SOIL 1437 0.73 0.3915 
SHADE 1 1437 1.54 0.2155 
SOIL*SHADE 1437 18.45 <.0001* 
TIME*SHADE 1 1437 1.52 0.218 
TIME*SOIL 1437 0.36 0.5492 
TIME*SOIL*SHADE 1 1437 20.8 <.0001* 
*Means significantly different (?=0.05) 
Neither the shade treatment nor soil condition was statistically significant with 
respect to height growth over the intervals, in overall tests, or in years evaluated.  An 
interaction between the shade treatment and soil condition was significant.  The 
interaction was not significant for the 1969 growth interval, but it was for 1970.  The 
significant interaction for the total growth over the 2 years showed that mean height 
growth for trees on the Wagram soil with no shade treatment and trees on the Lucy soil 
with the shade treatment was significantly greater than mean height growth for trees on 
the Wagram soil with the shading treatment.  When isolating 1970 growth, the same 
general interaction was occurring, but mean height growth for trees on the Lucy soil with 
no shade treatment was significantly lower instead.  From the overall interaction results, 
it looks like that shade significantly affected mean height growth on the Wagram soils.  
For the Lucy soil there was an opposite trend for 1970 growth where shaded trees on 
average grew significantly better than non shaded trees.  The lack of an overall 
significant shading treatment might be due to the percent of shading caused by the 
cheesecloth or the height it was placed.  The cheesecloth was suspended 1 meter above 
 
 
 
 33
the terminal bud, which allowed the tree to receive lateral rays of sunlight during the 
early morning and late afternoon.  This along with the low percent of sunlight reduced by 
the cheesecloth did not seem to change the growth rates of the shade trees in comparison 
to those that were not shaded at all.   
5.2  Flushes 
 
In 1969 all trees flushed 5 times with 62.5% starting a sixth bud and 5% having a 
sixth flush.  In 1970, all trees flushed 4 times, 92% had a fifth flush, and 46% had a sixth 
flush.  Figures 5.2.1 and 5.2.2 show the general form of mean flushes by year.  It is clear 
that more than one flush was growing at the same time.  Some trees start flushing later 
than others causing a dip in the curves as seen in Figure 5.2.2 with the fifth flush.  As 
expected, the first flush was consistently the largest for both years and formed an 
approximate sigmoid curve, but the rest of the flushes varied from a sigmoid shape (Allen 
and Scarbough 1969).  The first flush contained 39% of the total year?s growth in 1969 
and 35% in 1970, which is consistent with what Mudano et al. (1992) and Kozlowski and 
Pallary (1997) found in other pines.  The peak growth rate of the first flush of both years 
occurred during April but continued at a slower rate until May.  The growth observed in 
Figures 5.2.1 and 5.2.2 show continuous growth of flushes.  Growth of 2 flushes at the 
same time during the growing season was seen in this study, which has been commonly 
seen in southern pines (Allen and Scarbough 1969, Byram and Doolittle 1950, Griffing 
and Elam 1971, Mudano et al. 1992, and Tepper 1963).   
 
 
 
 
 
 
Figure 5.2.1  Average flush lengths of the sampled trees during 1969 
 
 34
 
 
 
 
Figure 5.2.2  Average flush lengths of sample trees during 1970   
 
 35
 
 
 
 36
This contrasts the consecutive growth patterns that Eggler (1961) reported.  In the case of 
this study, the next flush did seem to start growing when the previous flush was 
approaching its peak, but there was overlap between the two.  At no time were more than 
2 flushes growing at the same time.  The amount of growth achieved by each flush was 
consistently less than the previous flush. 
Growth for each flush was tested for significant differences with respect to the 
shading treatment and soil condition using the same procedures outlined in Chapter 5.1.   
None of the 1969 flushes were significantly different with respect to the shade treatment 
or soil condition.  Flush 5 in 1970 was significantly different with respect to soil type.  
Trees on the Wagram soils had significantly larger fifth flushes than those on the Lucy 
soils.  This finding is not supported by any of the other tests with height growth.    The 
shading treatment and soil condition were not significant for any of the other 1970 
flushes. 
Relationships between the lengths of buds and flush lengths were tested with the 
1969 and 1970 data.  Table 5.2.1 shows the regression equations for initial bud sets for 
each year and the length of the first flush during 1969 and 1970. The lengths of buds 
prior to the growing season have been found to be significantly correlated with the length 
of the first flush.  Bengston et al. (1967) also found a significant correlation with slash 
pine reporting an r of 0.97.  In 1969 and 1970, 21.33% and 30.26%, respectively, of the 
variability in the first internode of growth was explained by the length of the bud set prior 
to the growing season.  Boyer (1970) found a similar relationship in loblolly pine 
seedlings and saplings reporting the 76% and 78%, respectively, of the variability in the 
first internode could be explained by the bud set prior to the growing season.   The 
 
 
 
 37
relationships of the sampled longleaf pine seem to not be as strong as those of work done 
in slash and loblolly pine.  This leads to the idea that the environment during the growing 
season has more of an effect on the first flush growth than during the bud set of the 
previous year. 
Table 5.2.1  Bud and flush regressions of the sampled trees for 1969 and 1970 
 
Year Regression Equation R
2
 
1969 Flush1 = 2.73Bud69 + 30.96 0.2133 
 
1970   Flush1 = 3.61Bud70 + 25.582 0.3026 
Bud(Year) = length of the bud set prior to the growing season for that year, 
Flush1 = length of first flush,  
 
Looking further into the relationships between flushes, correlations between 
flushes were also seen. Flushes 3, 4, and 5 were found to be significantly correlated with 
each other in 1969 and again in 1970.  Flush 5 was also significantly correlated with flush 
6 in 1969 only.  Allen and Scarbrough (1970) also found successive flushes to be 
significantly correlated in longleaf pine, which they linked to similarities in internode 
elongation.  The correlation of flushes seems to occur during the middle and toward the 
end of the growing seasons instead of the beginning and ending.  Growth is at its highest 
during the first flush of the season.  This is because the tree is using all of its stored 
nutrients and soil moisture from the dormant season to grow (Kienholz 1941).  As the 
tree continues to have flushes of growth and reaches these dry summer months, growth 
rates are continuing to slow down.  During this time, separate flushes are growing more 
concurrently than in the beginning or ending.  The final flush of the growing season is 
setting the bud for the next growing season and is less likely to be correlated with the 
 
 
 
 38
other flushes in this growing season, but will have a relationship with the first flush of the 
next growing season. 
5.3  Needle Lengths 
 Needle lengths were measured with height growth from April 28 through October 
14 of 1970.  Needle length measurements were not measured during the 1969 growing 
season during the shading treatment. Therefore, only residual effects can be tested. The 
overall average needle growth during this interval was 31.65 cm.  Table 5.3.1 shows 
mean needle growth by soil and shade combinations, and Figure 5.3.1 shows the average 
needle lengths over the growing season.   Needles grew throughout the growing season 
following similar patterns of height growth.  Growth rates were highest at the beginning 
of the growing season during the April and May and continued to grow until reaching a 
peak and then leveling off in early August.  The following table shows average needle 
growth by soil, shade, and shade/soil combinations.  There seemed to be no significant 
differences between shading treatments, soil condition or significant interactions at the 
0.05 level of significance.     
 
 
 
 
 
 
 
 
 
 
 
 39
Table 5.3.1  Average needle growth of the sampled trees during 1970  
 
Treatment/Condition Average Needle Growth (cm) 
No Shade 32.45 
Shade 30.81 
Lucy 29.79 
Wagram 32.47 
Shade/Soil Combination Average Needle Growth (cm) 
NL 31.62 
NW 33.27 
SL 29.87 
SW 31.66 
*Means significantly different (?=0.05) 
NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5.3.1  Average needle lengths of the sampled trees during 1970 
 40
 
 
 
 41
5.4  1969 through 1970 Diameter Growth 
Evaluating differences in diameter growth over the different intervals is necessary 
to see if the recorded environmental conditions and treatments are affecting diameter 
growth.  For diameter growth, each year needs to be individually analyzed because of a 
calibration issue in January 1970.  See Appendix 5.4.1 for a full description of the error 
and for additional statistical procedures.  Tree 27 was in the early stages of mortality and 
caused a significant problem with the analyses.  It was removed where appropriate.  
Statistical procedures used to evaluate height growth will be followed.  
Monthly diameters were measured in 1969 from March 5 through December 24.  
Table 5.4.1 shows averages of diameter growth by shade treatment and soil type for 
1969.  A significant difference between trees on the Lucy loamy sand and the Wagram 
loamy sand was observed from testing overall growth from March of 1969 to December 
of 1969, shown in Table 5.4.2  All tests were conducted at the 0.05 level of significance.  
Table 5.4.1  Average diameter growth during 1969 monthly  
from March 1969 to December 1969  
 
Treatment/Condition Average Diameter Growth (cm) 
No Shade 1.37 a 
Shade 1.25 a 
Lucy 1.57 b 
Wagram  1.03 c 
Shade/Soil Combination Average Diameter Growth (cm) 
NL 1.50 d 
SL 1.36 d e  
NW 1.11 f e 
SW 1.06 f 
Means with the same letter are not significantly different (?=0.05) 
NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
 
 
 42
Table 5.4.2   Testing overall diameter growth over monthly interval  
from March 1969 to December 1969 using GLM procedure  
 
SOURCE DF Type III SS Mean Square F Value Pr > F 
SOIL 1 1.730625 1.730625 15.35 0.0004* 
SHADE 1 0.04866453 0.04866453 0.7 0.2674 
SOIL*SHADE 1 0.00038345 0.00038345 0.01 0.5903 
*Means significantly different (?=0.05) 
Diameters were measured in 1970 from January through December.  Table 5.4.3 
shows averages of diameter growth during the 1970 monthly interval by shade treatment 
and soil type for 1970.  Testing within group differences, soil type was significant, and 
mean diameter growth was nonsignficantly larger for the trees with the shade treatment 
on both soils.  
Table 5.4.3  Average diameter growth during 1970 monthly interval  
from January 1970 to December 1970  
 
Treatment/Condition Average Diameter Growth (cm) 
No Shade 1.49 a 
Shade 1.56 a 
Lucy 1.71 b 
Wagram  1.34 c 
Shade/Soil Combination Average Diameter Growth (cm) 
SL 1.76 d 
NL 1.67 
SW 1.38 e 
NW 1.30 
*Means with the same letter are not significantly different (?=0.05) 
NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
 
 
 
 
 
 
 43
A significant difference between trees on the Lucy loamy sand and the Wagram 
loamy sand was observed for overall growth during the 1970 monthly interval (Table 
5.4.4). 
Table 5.4.4  Testing overall diameter growth over monthly interval from January 1970 to 
December 1970 using GLM procedure  
 
SOURCE DF Type III SS Mean Square F Value Pr > F 
SOIL 1 1.19000075 1.19000075 17.18 0.0002* 
SHADE 1 0.04866453 0.04866453 0.7 0.4076  
SOIL*SHADE 1 0.00038345 0.00038345 0.01 0.9411  
*Means significantly different (?=0.05) 
Monthly diameter growth was explored for potential significant differences using 
the repeated measurements design.  Initially the data were combined into 22 months 
ranging from March of 1969 to December of 1970.  Tree 27 was also left in the dataset.  
Table 5.4.5 shows the results of this test.  
Table 5.4.5  Testing monthly diameter growth from March 1969 to  
December 1970 using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 1075 7162.66 <.0001 
SOIL 1075 107.43 <.0001* 
SHADE 1 1075 2.89 0.0893 
SOIL*SHADE 1 1075 1.62 0.2028 
TIME*SHADE 1 1075 2.38 0.1232 
TIME*SOIL 1 1075 161.67 <.0001* 
TIME*SOIL*SHADE 1 1075 0.47 0.4921 
*Means significantly different (?=0.05) 
This test reinforces the above analyses showing that there is a significant difference 
between trees with regard to soil type.  The same analysis was conducted with cumulative 
growth and with lagging growth by 1 and 2 weeks.  The results still only showed a 
 
 
 
 44
significant difference in relation to soil type.  However, removing tree 27 and running the 
same analysis did change the results. The output is in Table 5.4.6. 
Table 5.4.6  Testing monthly diameter growth from March 1969 to December 1970 
without tree 27 using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 1048 7066.52 <.0001* 
SOIL 1048 96.42 <.0001* 
SHADE 1 1048 4.85 0.0278* 
SOIL*SHADE 1048 3.16 0.0757 
TIME*SHADE 1 1048 5.07 0.0245* 
TIME*SOIL 1048 138.91 <.0001* 
TIME*SOIL*SHADE 1 1048 1.99 0.1589 
*Means significantly different (?=0.05) 
After tree 27 was removed, the shade treatment became significant.  Dividing the 
total dataset into a dataset for each year also changed the results.  Also, the months of 
January and February of 1970 were removed from the 1970 dataset due to an error in the 
original datasheet and to make the 1969 and 1970 datasets over the same span in time.  
Below are the results of the repeated measures test for each year in Tables 5.4.7 and 
5.4.8. 
Table 5.4.7  Testing 1969 diameter growth over the monthly interval from March to 
December without tree 27 using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 464 3557.67 <.0001* 
SOIL 464 48.72 <.0001* 
SHADE 1 464 3.52 0.0612 
SOIL*SHADE 464 2.47 0.1170 
TIME*SHADE 1 464 4.33 0.0380* 
TIME*SOIL 464 55.84 <.0001* 
TIME*SOIL*SHADE 1 464 1.61 0.2045 
*Means significantly different (?=0.05) 
 
 
 
 45
The results from the 1969 analysis show a significant difference for soil type and a 
significant interaction between time and shade.  The shade treatment is not significant at 
this level of significance.  The results from the 1970 analysis show a significant 
difference for soil type and the shade treatment.   
Table 5.4.8  Testing 1970 diameter growth over the monthly interval from March to 
December without tree 27 using repeated measures procedure  
 
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 541 8647.28 <.0001* 
SOIL 541 69.90 <.0001* 
SHADE 1 541 15.03 0.0001* 
SOIL*SHADE 541 0.53 0.4652 
TIME*SHADE 1 541 16.46 <.0000* 
TIME*SOIL 541 97.43 <.0001* 
TIME*SOIL*SHADE 1 541 0.20 0.6524 
*Means significantly different (?=0.05) 
The biweekly intervals from 1969 and 1970 were tested next. The results can be 
found in Tables 5.4.9 and 5.4.10, respectively.  There were no significant differences or 
interactions in diameter growth with respect to soil condition or shade treatment for the 
1969 interval.   The 1970 biweekly interval shows a significant soil shade interaction, but 
neither the shade treatment nor soil condition is statistically significant.    
Table 5.4.9  Testing diameter growth over biweekly interval from March 1969 to 
December 1969 using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 871 1676.82 <.0001* 
SOIL 871 0.39 0.5302 
SHADE 1 871 0.03 0.8733 
SOIL*SHADE 1 871 0.26 0.6113 
TIME*SHADE 1 871 0.02 0.8774 
TIME*SOIL 871 0.83 0.3634 
TIME*SOIL*SHADE 1 871 0.43 0.511 
*Means significantly different (?=0.05) 
 
 
 
 46
 
Table 5.4.10  Testing diameter growth over biweekly interval from March 1970 to 
December 1970 using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 864 1765.95 <.0001* 
SOIL 864 0.6 0.4373 
SHADE 1 864 0.59 0.4443 
SOIL*SHADE 1 864 11.89 0.0006* 
TIME*SHADE 1 864 0.57 0.4512 
TIME*SOIL 864 0.26 0.6078 
TIME*SOIL*SHADE 1 864 13.58 0.0002* 
*Means significantly different (?=0.05) 
Tables 5.4.11 and 5.4.12 contain the results to the repeated measures tests for the 
1969 and 1970 weekly intervals, respectively.  Both tests show significant differences 
between soil types and shade treatments.  There are also significant interactions between 
soil conditions and time and also the shade treatment and time, but there was not a 
significant interaction between the soil condition and shade treatment. 
Table 5.4.11  Testing diameter growth over 1969 weekly interval from March to 
October using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 1356 13699.4 <.0001* 
SOIL 1356 151.81 <.0001* 
SHADE 1 1356 18.79 <.0001* 
SOIL*SHADE 1 1356 3.17 0.0754 
TIME*SHADE 1 1356 25.33 <.0001* 
TIME*SOIL 1356 190.23 <.0001* 
TIME*SOIL*SHADE 1 1356 1.43 0.2318 
*Means significantly different (?=0.05) 
 
 
 
 
 
 
 
 
 
 
 
 47
Table 5.4.12  Testing weekly diameter growth over 1970 weekly interval from March to 
 November using repeated measures procedure  
 
EFFECT Num DF Den DF F Value Pr > F 
TIME 1 1636 16574.80 <.0001* 
SOIL 1636 173.13 <.0001* 
SHADE 1 1636 17.96 <.0001* 
SOIL*SHADE 1 1636 0.01 0.9436 
TIME*SHADE 1 1636 20.13 <.0001* 
TIME*SOIL 1636 267.03 <.0001* 
TIME*SOIL*SHADE 1 1636 0.53 0.4667 
*Means significantly different (?=0.05) 
 Diameter growth for trees on the Lucy loamy sands was significantly greater than 
diameter growth for trees on the Wagram loamy sands across all intervals except for the 
biweekly datasets. The reason for better growth on the Lucy loamy sands could possibly 
be that longleaf pine grows better on drier soils due to potentially less ground layer 
competition.  If the root systems were deep enough to reach the B-horizon there might be 
a moisture gradient caused by the higher clay content at shallower depths than the 
Wagram loamy sands.  The significant differences in diameter growth by soil type may 
also be due to competition of surrounding trees instead of soil differences, but without 
more detailed data about surrounding tree and vegetation this concept cannot be explored 
further. 
Repeated measures analyses also showed a significant shade treatment for 
diameter growth for monthly and weekly intervals in 1969 and 1970 when tree 27 was 
removed.  However, adjusting weekly interval tests for autocorrelation did affect the 
analyses, and the normality assumption was suspect for the weekly 1970 test.  
Calculating the overall means for these intervals, the shade treatment was not significant, 
 
 
 
 48
but testing growth within each soil type during 1970 showed that shaded trees grew less 
than non shaded trees.     
5.5  Growing Seasons 
Average daily temperature during the 1969 growing season was 28
?
C.  The total 
amount of precipitation during this time was 131 cm.  During the 1970 growing season, 
there was a total of 155 cm of precipitation with an average daily temperature of 29
?
C. 
Table 5.5.1 contains the percent of annual average height and diameter growth for 
each month during the 1969 and 1970 growing seasons.   Figure 5.5.1, provides plots of 
the percent of annual average height and diameter growth by month for both growing 
seasons.  Data recorded during both seasons showed very similar cyclic patterns for 
height and diameter growth during the two year data.  Height growth seems to begin in 
March, peak in April, and decline until ceasing in October and November.  Over 90% of 
the year?s height growth was completed during the month of August of both years.  
Mudano et al. (1992) reported that 90% of total shoot growth for loblolly pine was 
completed by the middle of August, and growth ceased in early November.   A similar 
growing season was also reported by Allen and Scarbrough (1969) for longleaf pine.  The 
month of April was the peak of the growing season for height growth, representing over 
30% of the total growth each year.  Mudano et al. (1992) and Kozlowski and Pallardy 
(1997b) also saw a peak of about 30% of annual height grow during the month of April.   
Diameter growth had similar patterns to height growth starting in March and 
peaking in April or May.  However, diameter growth did not consistently decline until 
October as height growth did.  Diameter growth was somewhat variable from June to 
October and even persisted into the winter months, which accounted for less than 5% of 
 
 
 
 49
the average diameter growth between November and December.  Over 90% of the years 
average diameter growth was completed by September or October for both years.  This 
shows that height growth is ceasing about a month before diameter growth for these two 
growing seasons.  Diameter growth is potentially more susceptible to changes in the 
environment than height growth as seen during the growing season and throughout the 
year (Kozlowski and Ward 1961).  Months following low precipitation seem the greatly 
affect percent diameter growth.  As seen in July 1969 and May 1970, percent diameter 
following low months of precipitation was reduced by half.   
 
Table 5.5.1   Percent of average annual height and  
diameter growth  by month from 1969 through 1970  
 
Year Month 
Average Height 
 Growth (%) 
Average Diameter 
 Growth (%) 
1969 March 3.30 11.02 
  April 31.77 18.07 
  May 24.85 18.80 
  June 17.26 7.67 
  July 11.10 12.98 
  August 8.21 12.09 
  September 2.46 12.42 
  October 0.80 1.96 
  November 0.23 3.17 
  December 0.03 1.82 
1970 January 0.00 0.00 
  February 0.51 1.52 
  March 2.31 5.53 
  April 31.95 20.04 
  May 17.95 9.78 
  June 16.82 17.01 
  July 13.86 10.94 
  August 10.98 14.85 
  September 3.17 8.97 
  October 1.63 9.37 
  November 0.74 1.44 
  December 0.09 0.55 
 
 
 
 
 
Figure 5.5.1  Percent of average monthly height and diameter growth plotted by each 
month data were collected in 1969 and 1970  
 
 50
 
 
 
 51
5.6  Diurnal and Nocturnal Growth 
The diurnal period between the morning and evening measurements was about 8 
hours, and the nocturnal period between evening measurements and the following 
morning measurements was about 16 hours (Boyer 1970).  Diurnal growth is the growth 
that occurred between the initial morning measurement and the initial afternoon 
measurement.  The nocturnal growth refers to the growth occurring between the initial 
afternoon measurement and the initial morning measurement.  The 24-hour growth refers 
to the growth occurring from the initial morning measurement until the initial morning 
measurement the following day. 
Specimens were measured from the morning of April 8, 1969 until the morning of 
April 28, 1969.  This resulted in 19 diurnal, nocturnal, and 24-hour measurements.  
During the 19 days the trees were measured in 1969, the average daily temperature was 
26
?
C.  The average temperature during the night was 12
?
C. Average height growth over 
the 19 days was 21.64 cm.  About 35% of the growth occurred during the diurnal period, 
while about 65% occurred during the nocturnal period.   
During the 16 days the trees were measured in April 1970, the average daily 
temperature was 28
?
C, while the average nightly temperature was 14
?
C.  The total 
amounts of precipitation during these measurement periods in 1969 and 1970 were 10.16 
cm and 11.43 cm, respectively.  During the 16 days the trees were measured in 1970, 
average height growth was 20.44 cm.  About 25% of the growth occurred during the 
diurnal period, and about 75% of the growth occurred during the nocturnal period.  The 
large percentage of growth occurring during the nocturnal periods is possibly due to the 
 
 
 
 52
time in which the trees were measured. The trees were also receiving light, but just not 
from directly above.  
 There were originally 20 measurement dates in 1970 instead of 16. The reason 
for the difference is that, on the 17th measurement, a new bud formed.  There were four 
measurements, including diurnal, nocturnal, and 24-hour, that included the new bud.  
These measurements were removed from the dataset. 
There was not a statistically significant difference between the soil types or the 
shade treatment for average diurnal or nocturnal growth during the measurement periods 
in April 1969 or 1970.  Procedures from Appendix 5.1.1 for analyzing mean differences 
in growth over numerous intervals were used.  To evaluate how the measured 
environmental variables influenced height growth during the peak of the growing season, 
a regression analysis was conducted.  Since height growth was measured in intervals of 
diurnal, nocturnal, and total day (24-hour), three sets of regression analyses were 
conducted on the datasets from each year.  Table 5.6.1 is a list of names and brief 
descriptions of variables used in the regressions.  Tables 5.6.2, 5.6.3, and 5.6.4 contain 
the significant single variable models developed from the regression analyses for 1969 
mean diurnal, nocturnal, and 24-hour growth, respectively.  Tables 5.6.5, 5.6.6, and 5.6.7 
contain the single variable models developed from the regression analyses for 1970 mean 
diurnal, nocturnal, and 24-hour growth, respectively.  Notice that no single variable 
consistently dominated all intervals.   
 During the peak of the growing season, temperature and precipitation were the 
most influential variables for height growth.  As Boyer (1970) found with loblolly pine, 
temperature and degree-hour accumulations seem to explain more variability in nocturnal 
 
 
 
 53
and 24-hour growth than any of the other variables.   All significant variables in the 1969 
diurnal analyses were lagged variables.  This shows that growth during the diurnal period 
is influenced by environmental conditions from the day before.  Precipitation from the 
day before explained about 58% of the variation in diurnal height growth, which was 
greater than any other variable.  Analyses from the 1970 diurnal data showed measures of 
temperature as better explanatory variables for mean diurnal height growth.  Degree-hour 
accumulations, greater than 40? F (4.4 ?C), explained about 46% of the variation in 
diurnal height growth.  However, the degree-hour accumulations greater than 45? F (7.2? 
C) and 50? F (10? C) were also significant and explained about the same amount of 
variability in diurnal height growth in 1970.  Lagged vapor pressure deficit was 
significant for both the 1969 and 1970 data. 
Temperature was the most influential variable in explaining the variation in 
nocturnal and 24-hour growth rates.  Degree-hour accumulations above 55? F (12.8 ?C) 
and 45? F for the 24-hour periods influenced mean nocturnal growth more than any other 
selected environmental variables for 1969 and 1970, respectively.   Degree-hour 
accumulations above 40? to 55? F for the 24-hour periods seemed to have the greatest 
influence on mean 24-hour growth.  However, mean 24-hour temperature explained 
about the same amount of the variability in mean 24-hour growth for both years as 
degree-hour accumulations did.  The results from mean diurnal, nocturnal, and 24-hour 
growth from both years consistently show temperature as the most influential 
environmental variable.    
 
 
 
 
 
 
 
 54
Table 5.6.1  Diurnal and nocturnal growth regression keywords 
Diurnal Interval 
HGDR  mean height growth rate (cm/hour) 
TIMED time interval (0.1hour) 
TMAXD maximum temperature (?C) 
TMIND minimum temperature (?C) 
TMEAND mean temperature (?C) 
WINDD wind speed (total miles) 
RHxDay relative humidity (percent) 
VPDD  vapor pressure deficit (mm Hg) 
D40D  degree-hours above 40? F (4.4? C) 
D45D  degree-hours above 45? F (7.2? C) 
D50D  degree-hours above 50? F (10? C) 
D55D  degree-hours above 55? F (12.8? C) 
Nocturnal Interval 
HGNR  mean height growth rate (cm/hour) 
TIMEN time interval (0.1 hour) 
TMAXN maximum temperature (?C) 
TMINN minimum temperature (?C) 
TMEANN mean temperature (?C) 
WINDN wind speed (total miles) 
RHxNight relative humidity (percent) 
VPDN  vapor pressure deficit (mm Hg) 
D40N  degree-hours above 40? F (4.4? C) 
D45N  degree-hours above 45? F (7.2? C) 
D50N  degree-hours above 50? F (10? C) 
D55N  degree-hours above 55? F (12.8? C) 
24 hour Interval 
H24R  mean growth rate (cm/hour) 
TIME24 time interval (0.1 hour) 
MAXT  maximum temperature (?C) 
MINT  minimum temperature (?C) 
TDIFF  difference between MAXT and MINT (?C) 
TMEAN24 mean temperature during (?C) 
PREC  total daily precipitation (cm) 
WINDT wind speed (total miles) 
RHx24Hr relative humidity (percent) 
VPD24 vapor pressure deficit (mm Hg) 
SRO  solar radiation open (langleys cal/cm3) 
SRS  solar radiation shaded (langleys cal/cm3) 
D40T  degree-hours above 40? F (4.4? C) 
D45T  degree-hours above 45? F (7.2? C) 
D50T  degree-hours above 50? F (10? C) 
D55T  degree-hours above 55? F (12.8? C) 
L# prefix to any variable means values have been lagged # of intervals 
 
 
 
 55
Table 5.6.2 1969 mean diurnal growth rate (HDR) regressions 
 
Independent Variable Equation R
2
Adj-R
2
L1PREC HDR = 0.05534 - 0.00769 L1PREC 0.5795 0.5532 
L1VPDD HDR = 0.3918 + 0.3283L1 VPDD 0.3296 0.2877 
L1RHX24 HDR = 0.6843 - 0.00029562 L1RHX24 0.3171 0.2245 
L1TMAXD HDR = -0.0487 + 0.00223 L1MAXD 0.2731 0.2276 
L1VPD24 HDR = 0.3971 + 0.08709 L1VPD24 0.2669 0.2211 
L1MAXT HDR = -0.00050303 + 0.00202 L1MAXT 0.2103 0.1609 
L1SRO HDR = 0.04072+ 0.0002178 L1SRO 0.2062 0.1566 
 
 
 
Table 5.6.3 1969 mean nocturnal growth rate (HNR) regressions 
 
Independent Variable Equation R
2
Adj-R
2
D55T HNR = 0.01368 + 0.00010629 D55T 0.5623 0.5365 
D40T HNR = 0.00082248 + 0.00007421 D40T 0.5262 0.4983 
D50T HNR = 0.01352 + 0.00008531 D50T 0.5258 0.4979 
D45T HNR = 0.00811 + 0.00007705 D45T 0.5257 0.4978 
TMEAN24 HNR = -0.01128 + 0.00308 TMEAN24 0.5076 0.3224 
L1VPDN HNR = 0.05670 - 0.17396 L1VPDN 0.3867 0.3484 
D40N HNR = 0.02464 + 0.00006372 D40N 0.3646 0.3272 
TMEANN HNR = -0.00758 + 0.00231 TMEANN 0.3601 0.3224 
D45N HNR = 0.02900 + 0.00006576 D45N 0.3563 0.3185 
D55N HNR = 0.3448 + 0.00009185 D55N 0.3551 0.3171 
D50N HNR = 0.03239 + 0.00007247 D50N 0.3424 0.3037 
L1RHXN HNR = -0.01841 + 0.00073727 L1RHXN 0.3392 0.2979 
TMINN HNR = 0.03143 + 0.00115 TMINN 0.3044 0.2635 
TMIND HNR = 0.00147 + 0.00224 TMIND 0.2624 0.2191 
L2TMIND HNR = 0.09174 - 0.0024 L2TMIND 0.2506 0.2006 
TMEAND HNR = -0.00662 + 0.00228 TMEAND 0.2334 0.1884 
L1VPD24 HNR = 0.05672 - 0.07690 L1VPD24 0.2277 0.1795 
L1RHX24 HNR = 0.01263 + 0.00043405 L1RHX24 0.2230 0.1744 
D55D HNR = 0.02266 + 0.00015566 D55D 0.2217 0.1759 
D50D HNR = 0.01696 + 0.00015336 D50D 0.2207 0.1748 
D45D HNR = 0.01240 + 0.00014641 D45T 0.2056 0.1588 
 
 
 
 
 
 
 
 
 
 
 
 
 56
Table 5.6.4  1969 mean 24 hour growth rate (H24R) regressions 
 
Independent 
Variable Equation R
2
Adj-R
2
D55T H24R = 0.02680 + 0.00007391 D55T 0.3977 0.3622
D50T H24R = 0.0254 + 0.00005777 D50T 0.3527 0.3146
D40T H24R = 0.01696 + 0.0000500 D40T 0.3491 0.3111
D45T H24R = 0.02200 + 0.00005166 D45T 0.3457 0.3072
TMEAN24 H24R = 0.0088 + 0.00207 TMEAN24 0.3357 0.2966
L2TMIND H24R = 0.08638 - 0.00190 L2TMIND 0.2740 0.2256
TMIND H24R = 0.0186 + 0.0007544 TMIND 0.2625 0.2191
D55N H24R = 0.0392 + 0.00006456 D55N 0.2566 0.2129
D40N H24R = 0.03313 + 0.00004256 D40N 0.2379 0.1931
TMEANN H24R = 0.02697 + 0.00126 TMEANN 0.2355 0.1906
D45N H24R = 0.3609 + 0.00004378 D45N 0.2311 0.1858
D50N H24R = 0.3820 + 0.00004902 D50N 0.2291 0.1838
 
 
 
 
Table 5.6.5  1970 mean diurnal growth rate (HDR) regressions 
 
 
Independent 
Variable Equation R
2
Adj-R
2
 D40D HDR = -0.00277 + 0.00014822 D40D 0.458 0.4128 
 D45D HDR = 0.00338 + 0.00014713 D45D 0.449 0.4031 
 D50D HDR = 0.00927 + 0.00014713 D50D 0.449 0.4031 
 L1VPDN HDR = 0.2659 + 0.16156L1 VPDN 0.412 0.363 
                L1WINDN HDR = 0.03222 + 0.00083362   
  L1WINDN 0.385 0.3338 
 D55D HDR = 0.1784 + 0.00013458 D55D 0.3732 0.3210 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 57
 
Table 5.6.6  1970 mean nocturnal growth rate (HNR) regressions 
 
Independent Variable Equation R
2
Adj-R
2
D45T HNR = -0.04964 + 0.00019681 D45T 0.9696 0.9562 
D50T HNR = -0.03292 + 0.00020939 D50T 0.9617 0.9585 
D40T HNR = -0.07109 + 0.00019405 D40T 0.9590 0.9556 
TMEAN24 HNR = -0.11006 + 0.00848 TMEAN24 0.9471 0.9427 
D50N HNR = -0.00041073 + 0.00027755 D50N 0.9453 0.9407 
D55N HNR = 0.0127 + 0.0003173 D55N 0.9432 0.9384 
D55T HNR = -0.0172 + 0.0002330 D55T 0.9385 0.9334 
D40N HNR = -0.03069 + 0.00024429 D40N 0.9295 0.9236 
TMINN HNR = -0.00970 + 0.00504 TMINN 0.9226 0.9161 
D45N HNR = -0.01253 + 0.00024829 D45N 0.9215 0.9257 
TMEANN HNR = -0.06173 + 0.00698 TMEANN 0.9202 0.9140 
D40D HNR = -0.15141 + 0.000692 D40D 0.7869 0.7691 
D45D HNR = -0.12292 + 0.00068773 D45D 0.7733 0.7544 
D50D HNR = -0.09541 + 0.00068773 D50D 0.7733 0.7544 
TMEAND HNR = -0.19139 + 0.00979 TMEAND 0.7606 0.7406 
TMAXD HNR = -0.22491 + 0.01018 TMAXD 0.7083 0.6840 
D55D HNR = -0.05908 + 0.00064978 D55D 0.6858 0.6597 
TMIND HNR = -0.07181 + 0.00617 TMIND 0.6339 0.6034 
L2D55N HNR = 0.03415 + 0.00023597 L1D55N 0.5011 0.4512 
L1D55N HNR = 0.03221 + 0.00021267 L1D55N 0.4711 0.4232 
L1TMAXD HNR = -0.14964 + 0.00762 L1TMAXD 0.4593 0.4101 
L1WINDN HNR = 0.02303 + 0.00291 L1WINDN 0.3628 0.3097 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 58
 
Table 5.6.7  1970 mean 24-hour growth rate (H24R) regressions 
 
Independent Variable Equation R
2
Adj-R
2
D40T H24R = -0.04281 + 0.00014364 D40T 0.9480 0.9437
D45T H24R = -0.0269 + 0.00014563 D45T 0.9479 0.9436
D50T H24R = -0.01445 + 0.00015475 D50T 0.9476 0.9432
TMEAN24 H24R = -0.07160 + 0.00627 TMEAN24 0.9354 0.9300
D50N H24R = 0.01027 + 0.0002009 D50N 0.9275 0.9214
D55N H24R = 0.01947 + 0.00023309 D55N 0.9182 0.9114
D40N H24R = -0.0127 + 0.00018055 D40N 0.9161 0.9091
TMEANN H24R = -0.03566 + 0.00515 TMEANN 0.9054 0.8981
D40D H24R = -0.10305 + 0.00051479 D40D 0.7857 0.7679
D45D H24R = -0.08185 + 0.00051163 D45D 0.7721 0.7531
D50D H24R = -0.06138 + 0.00051143 D50D 0.7721 0.7531
TMEAND H24R = -0.13298 + 0.0729 TMEAND 0.7610 0.7411
TMAXD H24R = -0.15845 + 0.00760 TMAXD 0.7121 0.6881
TMIND H24R = -0.04292 + 0.00455 TMIND 0.6211 0.5895
L2D50N H24R = 0.02713 + 0.00015439 L2D50N 0.5612 0.5173
L1D50N H24R = 0.02616 + 0.00014190 L1D50N 0.5255 0.4823
L1WINDN H24R = 0.02599 + 0.00223 L1WINDN 0.3867 0.3356
RHX24 H24R = -0.06338 + 0.00167 RHX24 0.3555 0.3059
RHXN H24R = -0.09761 + 0.00181 RHXN 0.3168 0.2642
 
5.7  Height and Diameter Yearly Measurements 
 Height measurements were made for the 13 growing seasons from 1969 through 
1981, while diameter measurements are limited to the 1972 through 1981 growing 
seasons since the trees were measured at dbh during this period  
 Mortality and damage affected 6 trees during the remeasurement period.  Trees 1 
and 12 were reported as having broken tops during the 1979 measurement.  Tree 17 was 
identified as being dead in 1978, while tree 27 was identified as dead in 1973.  Tree 22 
was also reported as having a broken top in 1980.  Tree 33 was reported as having a 
broken top in 1973, but by 1975 a new leader emerged.  Tree 40 was identified as a 
leaning tree, but this did not seem to affect its growth during this interval.  No cause of 
 
 
 
 59
death or injury was recorded for the trees except for tree 17 which had a stem canker 
caused by fusiform rust (Cronartium quercumm f.sp fusiforme).  Data associated with 
these trees were removed from the data set where appropriate. 
 Average height growth over the 13 seasons of the entire study was 13.7 m.  Over 
the 10 growing seasons of the remeasurement period, average height growth was 10.3 m.  
Average diameter growth during the remeasurement period was 7.8 cm.  Appendix 5.7.1 
contains plots of height and diameter measurements and yearly growth by year. 
 An analysis of variance at the 0.05 level of significance was conducted to test for 
significant differences of heights and diameters with respect to soil type and the 
possibility of a shade treatment residual effect.  Appendix 5.7.1 contains the tests used in 
the analysis of height growth.  The same procedures for evaluating mean differences in 
Appendix 5.1.1 were followed.  All regression analyses followed the same procedures as 
Appendix 5.6.1.   
 There were no significant differences in height growth with respect to soil 
condition or residual effects from the shade treatment.  Table 5.7.1 is the effects table 
associated with the test of overall height growth over the entire study.  The normality 
assumption did not seem to be violated.  The QQ plot was acceptable and the Anderson-
Darling test resulted in a nonsignificant p-value.  Neither Barlett?s nor Levene?s test for 
homogeneity of variance were significant, indicating no violations of the assumption of 
constant variance.  Height growth during the remeasurement period was also tested, but 
there were no significant results. 
 
 
 
 
 
 
 
 60
 
 
Table 5.7.1  Testing overall height growth from 1969  
through 1981 using GLM procedure  
 
Source DF Type III SS Mean Square F Value Pr > F 
SOIL 1 4793.1668 4793.1668 0.78 0.3831 
SHADE 1 12315.851 12315.85055 2.01 0.1661 
SOIL*SHADE 1 1721.1232 1721.1232 0.28 0.5998 
*Means significantly different (?=0.05) 
 There was a significant difference in average diameter growth during the 
remeasurement period for soil type.  The average diameter growth for trees on the Lucy 
loamy sand was 8.8 cm and 6.3 cm on the Wagram loamy sand with a significant p-value 
of 0.0002.  The difference in mean yearly diameter growth seemed to increase throughout 
the yearly measurements interval as seen in Figure 5.7.1.  Diameter growth on the Lucy 
soil is possibly better because of the shallow B-horizon, which could provide roots in this 
horizon with more soil moisture due to higher clay content.  The effects table from this 
test can be found in Table 5.7.2.  There were no significant residual differences between 
shade treatments or significant interactions.   The QQ plot was acceptable and the 
Anderson-Darling test resulted in a nonsignificant p-value of  > 0.250.   However, both 
Barlett?s and  Levene?s tests for homogeneity of variance were significant, indicating 
violations of the assumption of constant variance.  Performing a log transformation of 
diameter growth solved the homogeneity of variance problem.  Testing the transformed 
diameter growth resulted in a significant difference in soil type.   A nonparametric 
Kruskal-Wallis test was used to test overall diameter growth with respect to soil 
condition.  The results were consistent with the results from Table 5.7.3.  The same 
 
 
 
 61
nonparametric test was used to test for shade treatment effects, but the shade treatment 
was not significant. 
 
Table 5.7.2  Testing overall diameter growth from 1972  
through 1981 using GLM procedure  
 
Source DF Type III SS Mean Square F Value Pr > F 
SOIL 1 56.355012 56.35501209 18.11 
                 
0.0002* 
SHADE 1 4.1262009 4.12620089 1.33  0.2583   
SOIL*SHADE 1 2.0410904 2.0410904 0.66 0.4242 
*Means significantly different (?=0.05) 
 
Table 5.7.3  Testing overall diameter growth from 1972  
through 1981 using nonparametric Kruskal-Wallis Test  
 
Chi-Square 12.7883
DF 1
Pr > Chi-Square 0.0003
*Means significantly different (?=0.05) 
 Table 5.7.4 contains the regression keywords for the yearly remeasurement 
regressions.  Table 5.7.5 contains yearly height and diameter regression results.  
Regression analyses showed temperature and precipitation explained the most variation 
in yearly height and diameter growth.   In the single variable equations, dormant season 
minimum temperature and precipitation seem to explain the most variability in yearly 
height growth.  Dormant season temperature difference and precipitation seem to be the 
best predictors for yearly diameter growth, with yearly diameter growth having a positive 
relationship with dormant season precipitation and a negative relationship with the 
difference between average maximum and minimum daily temperature.  Zahner (1956) 
and Shoulders and Tiarks (1980) both cite water as one of the most important resources 
 
 
 
 62
to a tree.  As the amount of precipitation during the dormant season increases, the amount 
of yearly height and diameter growth also increases.  Coile (1936) saw a positive 
correlation between longleaf pine growth and February through April rainfall.  A build up 
of soil moisture reserves during the dormant season aides a tree in peak photosynthesis 
during the beginning of the growing season.   
 
Table 5.7.4  Yearly remeasurement regression keywords  
 
Independent 
Variable Description 
HGT Yearly height growth (cm) 
DIA Yearly diameter growth (cm) 
MAXT Average maximum temperature during the year (?C) 
MINT Average minimum temperature during the year (?C) 
TDIFF Difference between MAXT and MINT (?C) 
PREC Sum of precipitation during the year (cm) 
GSMAXT Average maximum temperature during the growing season (?C) 
GSMINT Average minimum temperature during the growing season (?C) 
GSTDIFF Difference between GSMAXT and GSMINT (?C) 
GSPREC Sum of precipitation during the growing season (cm) 
DSMAXT Average maximum temperature during the dormant season (?C) 
DSMINT Average minimum temperature during the dormant season (?C) 
DSTIFF Difference between DSMAXT and DSMINT (?C) 
DSPREC Sum of precipitation during the dormant season (cm) 
Growing Season (March ?October), Dormant Season (November-February) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 63
Table 5.7.5  Yearly height and diameter regressions  
 
Yearly height growth (HGT) from 1969 through 1981  
MODEL R
2
ADJ-R
2
HGT = 71.9410 + 7.2793DSMINT 0.1681 0.1659 
HGT = 63.2694 + 0.8957DSPREC 0.1343 0.1320 
HGT = 280.112 - 12.8963TDIFF 0.0975 0.0951 
HGT = 21.3574 + 4.5378DSMAXT 0.0723 0.0699 
HGT = 253.657 - 10.855TDIFF 0.0376 0.0350 
HGT = 123.885 - 0.1894 GSPREC 0.0223 0.0197 
HGT = 60.1002 + 3.5218MINT 0.0149 0.0123 
  
HGT = 42.2710 + 6.3953DSMINT + 0.7534DSPREC 0.2607 0.2567 
HGT = -32.8265 + 0.9661DSPREC + 5.1655DSMAXT 0.2272 0.2230 
HGT = 366.687 - 0.4649GSPREC -15.2431GSTDIFF 0.0707 0.0657 
 
Yearly diameter growth (DIA) from 1972 through 1981 
MODEL R
2
ADJ-R
2
DIA = 3.4356 - 0.1951DSTDIFF 0.1716 0.1694 
DIA = 0.2535 + 0.0113DSPREC 0.1636 0.1614 
DIA = 3.6789 - 0.02106TDIFF 0.1083 0.1060 
DIA = 2.4162 - 0.1190GSTDIFF 0.0443 0.0418 
DIA = 0.6050 + 0.0356DSMINT 0.0305 0.0280 
DIA = 0.3870 + 0.0024PREC 0.0297 0.0272 
DIA = 1.5440 - 0.0302MAXT 0.0117 0.0092 
DIA = 1.7921 - 0.0351GSMAXT 0.0104 0.0078 
  
DIA = 2.1606 + 0.0066DSPREC - 0.1235DSTDIFF 0.2035 0.1994 
DIA = 0.0997 + 0.0114DSPREC + 0.0083DSMAXT 0.1654 0.1640 
DIA = 4.8875 - 0.0050GSPREC - 0.2565GSTDIFF 0.1026 0.1013 
 
 
 
 
 
 
 
 
 
 
Figure 5.7.1  Average diameter measurements by year and soil type  
from 1972 through 1981  
   
 64
 
 
 
 65
6. CONCLUSIONS 
 The desire to identify and develop a better understanding about the many 
relationships that exist between the environment and tree growth was the main purpose of 
this research effort.  To determine this, height and diameter growth were evaluated over 
numerous intervals during various portions of the growing and dormant seasons.  It was 
necessary to examine seasonal patterns of growth to understand how trees grow before 
attempting to determine what is affecting growth. 
 Both height and diameter growth displayed cyclic patterns during the two years 
the sample trees were measured intensively.  The growing season began in March and 
tended to peak in April or May.  Height growth occurred in flushes throughout the 
growing season with continuous growth of up to two flushes at one time.  The proceeding 
flush began as the previous flush?s growth rate slowed down.  Trees flushed at least five 
times in 1969 and four times in 1970.  During both years, over 30% of total height 
growth occurred in the month of April and in the first flush of the growing season.  
Height growth and flush lengths declined from this point on completing 90% of the 
year?s growth by the end of August and finally ceasing in early November.  Diameter 
growth was not as consistent as height growth.  Diameter growth was variable from June 
to October of both years and even persisted into the winter months.  However, 90% of the 
years total diameter growth was completed by September or October for both years.  The 
variation in diameter growth was not consistent over the two years. 
 
 
 
 66
 This potentially showed that diameter growth was being influenced by the environment 
more than height growth. 
 To evaluate height growth better, a more intense analysis was conducted during 
the April of each year, which was the peak of seasonal height growth.  Growth during this 
period was divided into diurnal and nocturnal growth.  The majority of the growth 
occurred during the nocturnal period accounting for 65% and 75% of total daily growth 
during 1969 and 1970, respectively.  The nocturnal measurements include growth before 
8:00 am and after 4:00 pm.  These are peak times for photosynthesis when solar radiation 
is down and moisture is more plentiful, which is why the amount of growth during the 
nocturnal growth period is dramatically larger than that of the diurnal period.  With a 
better understanding of height and diameter growth patterns, it is now possible to 
evaluate how the environment affects these patterns. 
Longleaf pine is commonly known to be very shade intolerant.  To determine how 
much light intensity affects the growth of longleaf pine, a shading treatment was 
randomly installed to half of the sampled trees.  The shading treatment resulted in a 
reduction in direct solar radiation of about 38% during the growing season of 1969.  The 
treatment seemed to have no significant effect on height growth of young longleaf pines 
over any of the intervals analyzed except for a significant effect on average height growth 
for trees on Wagram soils when looking at growth over the two year period.  This could 
be a residual effect from the shading treatment.  Weekly diameter growth of both 1969 
and 1970 were significant with respect to the shade treatment.  Mean diameter growth 
was also lower, however not significant, for shaded trees on both soil types.  The lack of 
an overall effect on height growth potentially occurred because of several reasons.  The 
 
 
 
 67
shading treatment was maintained at a meter or greater above the growing tip of each 
tree.  This allowed the terminal bud to receive latent rays of sunlight in the morning and 
evening which are the most important times for growth during the diurnal cycle.  
Duration is another key factor.  The treatment was only applied for one growing season.  
It may take a longer duration of shading to see an effect on growth.   
Soil type was another key factor that needed to be evaluated in determining a 
possible influence on growth.  The sample trees were on two different soil types and the 
two measures of growth were affected in different ways.  Soil type did not seem to have a 
significant effect on height growth except for the fifth flush of 1970, but there were 
differences in diameter growth.  Soil type significantly affected diameter growth over all 
intensities measured during the first two years of the study including the dormant season 
measurements.  Diameter growth for trees on the Lucy loamy sands was greater than 
diameter growth for trees on the Wagram loamy sands.  The difference in growth seemed 
to increase as the trees got older. This could possibly be due to a moisture gradient 
caused by the higher clay content and shallower B-horizon in the Lucy loamy sands.  The 
significant differences in mean diameter growth by soil type may also be due to 
competition of surrounding trees instead of soil differences, but without more detailed 
data about surrounding tree and vegetation this concept cannot be explored further. 
 Temperature and precipitation seemed to be the most influential environmental 
conditions in predicting height and diameter growth.  Measures of temperature and 
precipitation during the dormant season seem to explain most of the variability in yearly 
height and diameter growth.  Dormant season minimum temperature and dormant season 
precipitation seemed to explain more of the variability in yearly height growth than other 
 
 
 
 68
measured factors.  Dormant season temperature difference and precipitation seem to be 
the best predictors for yearly diameter growth, with yearly diameter growth having a 
positive relationship dormant season precipitation and a negative relationship with 
dormant season temperature difference. 
During the peak of the growing season, temperature and precipitation also seemed 
to be the most influential variables for mean height growth rates.  Mean diurnal growth 
rates seemed to be influenced more by temperature, vapor pressure deficit, and 
precipitation from the night or 24-hour period before.  Temperature and degree-hour 
accumulations for the 24-hour periods influenced mean nocturnal and mean 24-hour 
growth rates more than the other selected environmental variables.  Degree-hour 
accumulations above 45? and 55? F for the 24-hour periods seemed to have the greatest 
influence on mean diurnal growth rates and mean 24-hour growth rates.  This shows that 
a temperature was the most influential environmental variable affecting mean height 
growth during the peak of the growing season. 
 This study found temperature and precipitation to be the most influential 
environmental variables affecting the growth of the sampled young longleaf pine.  
However, there are still many unanswered questions about influences of the environment 
on the growth of young longleaf.  Understanding more about relationships between tree 
growth and the environment is very important to discovering more about patterns of tree 
growth.  To develop a better understanding of these relationships, a larger dataset is 
required to measure trees across different sites and over more time.  
 
 
 
 69
LITERATURE CITED 
 
Addington, R.N., R.J. Mitchell, R. Oren, and L.A. Donovan.  2004.  Stomatal sensitivity 
to vapor pressure deficit and its relationship to hydraulic conductance in Pinus 
palustris.  Tree Physiology.  24: 561-569. 
Allen, R.M. and N.M. Scarbrough.  1969.  Development of a year?s height growth in 
longleaf pine saplings.  USDA Forest Service, Southern Forest Experiment Station, 
Research Paper. SO-45:  13p. 
Allen, R.M. and N.M. Scarbrough.  1970.  Morphology and length correlated in terminal 
flushes of longleaf pine saplings.  USDA Forest Service, Southern Forest 
Experiment Station, Research Paper. SO-53:  15p.  
Barnett, J.P.  1989.  Shading reduces growth of longleaf and loblolly pine seedlings in 
containers.  Tree Planters Notes.  Winter: 23-26. 
Bengston, G.W., W.H. McGregor, and A.E. Squillace.  1967.  Phenology of terminal 
growth in slash pine:  some differences related to geographic seed source.  Forest 
Science. 13(4): 402-412. 
Boyer, W.D.  1970.  Shoot growth patterns of young loblolly pine.  Forest Science. 16(4): 
472-482.   
Boyer, W.D.  1976.  Thermal efficiency:  a possible determinant of height growth 
potential in young loblolly pine. Forest Science. 22(3): 279-282.
 
 
 70
Boyer, W.D. 1990. Longleaf pine. pp.405-412, In: Burns, R.M.; Honkala, B.H. (tech. 
coord.),   Silvics of North America, Volume I Conifers. U.S. Department of 
Agriculture, Forest Service Agriculture Handbook 654.  
Boyer, W.D., G.A. Ward, and J.S. Kush.  1997.  The Escambia Experimental Forest marks 
fifty years of research on the ecology and management of longleaf pine.  Southern 
Journal of Applied Forestry.  21(1): 47. 
Brown, C.L.  1964.  The seedling habit of longleaf pine.  Georgia Forest Research 
Council.  Report No. 10:  68p. 
Byram, G.M. and W.T. Doolittle.  1950.  A year of growth for a shortleaf pine.  Ecology.  
31(1): 27-35. 
Cody, R.P. and J.K. Smith.  1997.  Applied statistics and the SAS programming language 
4
th
 edition.   Upper Saddle River, NJ:  Prentice Hall.  445p. 
Coile, T.S.  1936.  The effect of rainfall and temperature on the annual radial growth of 
pine in the southern United States.  Ecological Monographs.  6(4): 533-562. 
Eggler, W.A.  1961.  Stem elongation and time of cone initiation in southern pines.  
Forest Science.  7(2): 149-158. 
Fritts, H.C.  1958.  An analysis of radial growth of beech in a central Ohio forest during 
1954-1955.  Ecology.  39(4):  705-719. 
Fritts, H.C.  1960.  Multiple regression analysis of radial growth in individual trees.  
Forest Science. 6(4): 334-349. 
Frost, C.C.   1993.  Four centuries of changing landscape patterns in the longleaf pine 
ecosystem.  In:  Hermann, S.H., editor.  Proceedings of the 18th Tall Timbers fire 
ecology conference.  The longleaf pine ecosystem:  ecology, restoration, and 
 
 
 71
management; 1991 May 30 ? June 2; Tallahassee, FL. Tallahassee(FL): Tall 
Timbers Research Station. 18: 139-158. 
Griffing, C.G. and W.W. Elam.  1971.  Height growth patterns of loblolly pine saplings.  
Forest Science.  17(1): 52-54. 
Jose, S., S. Merritt, and C.L. Ramsey.  2003.  Growth, nutrition, photosynthesis, and 
transpiration responses of longleaf pine seedlings to light, water, and nitrogen.  
Forest Ecology and Management. 180: 335-344. 
Kienholz, R. 1941.  Seasonal course of height growth in some hardwoods in Connecticut. 
Ecology.  22(3):  249-258 
Kozlowski, T.T.  1955.  Tree growth, action, and interaction of soil and other factors.  
Journal of Forestry.  53: 508-512. 
Kozlowski, T.T. and S.G. Pallardy. 1997a. Growth control in woody plants. San Diego, 
CA: Academic Press, Inc. 641p. 
Kozlowski, T.T. and S.G. Pallardy. 1997b. Physiology of woody plants second edition. 
San Diego, CA: Academic Press, Inc. 411p. 
Kozlowski, T.T. and R.C. Ward.  1961.  Shoot elongation characteristics of forest trees.  
Forest Science.  7(4):  357-368. 
Kramer, P.J.  1957.  Some effects of various combinations of day and night temperatures 
and photoperiod on the height growth of loblolly pine seedlings.  Forest Science. 
3(1): 45-55. 
Kramer, P.J. and J.S. Boyer.  1995.  Water relations of plants and soils. San Diego, 
CA:Academic Press, Inc. 495p. 
 
 
 72
Littell, R.C., G.A. Milliken, W.W. Stroup, and R.D. Wolfinger.  1996.  SAS System for 
Mixed Models.  SAS Institute Inc., Cary, North Carolina:  633p. 
Lodewick, J.E.  1930.  Effect of certain climatic factors on the diameter growth of 
longleaf pine in western Florida.  Journal of Agricultural Research.  41(5): 349-
363. 
Mattox, M.G., L.A. Duncan, H.B. Neal, W.B. Parker, and T.A. Walkup. 1975.  Soil 
survey of Escambia County, Alabama. United States Department of Agriculture, 
Soil Conservation. Service and Forest Service, in cooperation with the Alabama 
Department of Agriculture and Industries, and the Alabama Agricultural 
Experiment Station. 98p. 
McGuire, J.P., R.J. Mitchell, E.B. Moser, S.D. Pecot, D.H. Gjerstad, and C.W. Hedman.  
2001.  Gaps in a gappy forest:  plant resources, longleaf pine regeneration, and 
understory response to tree removal in longleaf pine savannas.  Canadian Journal 
of Forest Research.  31: 765-778.  
McLaughlin, S.B., S.D. Wullschleger, and M. Nosal.  2003.  Diurnal and seasonal 
changes in stem increment and water use by yellow poplar trees in response to 
environmental stress.  Tree Physiology. 23: 1125-1136. 
Mudano, J.E., H.L. Allen, and L.W. Kress. 1992.  Stem and foliage elongation of young 
loblolly pine as affected by ozone.  Forest Science. 38(2): 324-335. 
Neter, J., M.H. Kutner, C.J. Nachtsheim, and W. Wasserman. 1996.  Applied linear 
statistical models 4
th
 edition.  Boston, MA: McGraw-Hill:  1408p. 
Outcalt, K.W.  2000.  The longleaf pine ecosystem of the south.  Native Plants Journal.  
1: 43-44,47-53. 
 
 
 73
Palik, B.J., R.J. Mitchell, G. Houseal, and N. Pederson.  1997.  Effects of canopy 
structure on resource availability and seedling responses in a longleaf pine 
ecosystem.  Canadian Journal of Forest Research. 27: 1458-1464. 
Pessin, L.J.  1935.  Root habits of longleaf pine seedlings.  USDA Forest Service, 
Southern Forest Experiment Station, Occasional Paper.  43: 7p. 
Rai, R.B.  1995.  The impact of rainfall and temperature on the tree and stand growth of 
natural longleaf pine (Pinus Palustris Mill.).  Unpublished doctoral dissertation, 
Mississippi State University.  46p. 
Rayamajhi, J.N.  1996.  Productivity of natural stands of longleaf pine in relation to 
climatic factors.  Unpublished doctoral dissertation.  Auburn University.  177p. 
Ramsey, F.L., and D.W. Schafer.  2002.  The statistical sleuth a course in methods of data 
analysis second edition.  Duxbury, Pacific Grove, CA:  742p. 
Ramsey, C.L., S. Jose, B. Brecke, and S. Merritt.  2003.  Growth response of longleaf 
pine (Pinus palustris Mill.) seedlings to fertilization and herbaceous weed control 
in an old field in southern USA.  Forest Ecology and Management. 172(2-3): 281-
289. 
SAS Institute Inc., 2000-2004.  SAS 9.1.3 Help and Documentation, Cary, NC: SAS 
Institute Inc. 
SAS Institute Inc., 2003. Version 9.1, Cary, NC:  SAS Institute Inc. 
Shoulders, E. and A.E. Tiarks.  1980.  Predicting height and relative performance of 
major southern pines from rainfall, slope, and available soil moisture.  Forest 
Science. 26(3): 437-447. 
 
 
 74
Soil Survey Division, Natural Resources Conservation Service (2003), United States 
Department of Agriculture. Official Soil Series Descriptions [Online WWW]. 
Available URL:  " http://ortho.ftw.nrcs.usda.gov/osd/ " [Accessed 27 Oct 2003].  
Tepper, H.B.  1963.  Leader growth of young pitch and shortleaf pines.  Forest Science. 
9(3): 344-353. 
Treubig, R.J.  1960.  Relation of summer soil moisture deficiencies to annual radial 
increment of longleaf and loblolly pine.  Unpublished master of forestry thesis, 
Louisiana State Univeristy.  94p. 
Wahlenberg, W.G. 1946.  Longleaf pine:  its use, ecology, regeneration, protection, 
growth, and management.  Charles Lathrop Pack Forestry Foundation, 
Washington, DC:  429p. 
Yeh, H. and L.C. Wensel.  2000. The relationship between tree diameter growth and 
climate for coniferous species in northern California.  Canadian Journal of Forest 
Research. 30: 1463-1471. 
Zahner, R.  1956.  Evaluating summer water deficiencies.  USDA Forest Service, 
Southern Forest Experiment Station, Occasional Paper.  150: 18p. 
 
 
 
 
 
 
 
 
 
 75
APPENDICES 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Appendix 4.1.1 Soil core information from neutron probe tubes 
 
 
 
This is a chart of the soil core information from the study site.  It shows the breakdown of 
depths to each horizon.  
 76
 
 
 77
Appendix 4.4.1 Soil moisture plots and precipitation correlation output 
PREC   =  Precipitation 
PREC1 =  Precipitation lagged 1 interval 
PLAG2 =  Precipitation lagged 2 interval 
PLAG3 =  Precipitation lagged 3 interval 
PLAG4 =  Precipitation lagged 4 interval 
D##      =  Soil moisture at depth (##) in cm 
Pvalues less than the 0.05 level of significance are considered significant 
 
 
 PREC PREC1 PLAG2 PLAG3 PLAG4 D10 D25
  
PREC 1 0.18889 0.15449 0.12003 0.08549 0.02839 -0.0577
  <.0001 <.0001 0.0025 0.032 0.4762 0.1472
 632 632 631 630 629 632 632
  
PREC1 0.18889 1 0.86404 0.728 0.59186 0.51332 0.54091
 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
 632 640 639 638 637 640 640
 
      
Pearson Correlation Coefficients 
Prob> |r| under H0:  Rho=0 
Number of Observations 
  
      
 D50 D75 D100 D125 D150 D175  
       
PREC -0.1015 -0.1092 -0.061 -0.0473 -0.0441 -0.0383  
 0.0107 0.006 0.1258 0.2349 0.2683 0.336  
632 632 632 632 632 632  
        
PREC1 0.53745 0.51694 0.25576 0.16919 0.11441 0.07605  
 <.0001 <.0001 <.0001 <.0001 0.0038 0.0545  
640 640 640 640 640 640  
 
 
 78
Pearson Correlation Coefficients 
Prob> |r| under H0:  Rho=0 
Number of Observations 
        
 PREC PREC1 PLAG2 PLAG3 PLAG4 D10 D25
  
PLAG2 0.15449 0.86404 1 0.86401 0.72792 0.478 0.51292
 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
 631 639 639 638 637 639 639
  
PLAG3 0.12003 0.728 0.86401 1 0.86397 0.44597 0.48861
 0.0025 <.0001 <.0001 <.0001 <.0001 <.0001
 630 638 638 638 637 638 638
  
PLAG4 0.08549 0.59186 0.72792 0.86397 1 0.41484 0.46003
 0.032 <.0001 <.0001 <.0001 <.0001 <.0001
 629 637 637 637 637 637 637
  
D10 0.02839 0.51332 0.478 0.44597 0.41484 1 0.92691
 0.4762 <.0001 <.0001 <.0001 <.0001  <.0001
 632 640 639 638 637 640 640
  
D25 -0.0577 0.54091 0.51292 0.48861 0.46003 0.92691 1
 0.1472 <.0001 <.0001 <.0001 <.0001 <.0001 
 632 640 639 638 637 640 640
  
D50 -0.1015 0.53745 0.51392 0.49127 0.4703 0.7414 0.89453
 0.0107 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
 632 640 639 638 637 640 640
  
D75 -0.1092 0.51694 0.49429 0.47474 0.4575 0.6829 0.82711
 0.006 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
 632 640 639 638 637 640 640
  
D100 -0.061 0.25576 0.22431 0.21534 0.20683 0.4271 0.45079
 0.1258 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
 632 640 639 638 637 640 640
 
 
 
 
 
 
 
 79
 PREC PREC1 PLAG2 PLAG3 PLAG4 D10 D25
       
D125 -0.0473 0.16919 0.13596 0.12994 0.12776 0.34235 0.34809
 0.2349 <.0001 0.0006 0.001 0.0012 <.0001 <.0001
 632 640 639 638 637 640 640
  
D150 -0.0441 0.11441 0.07867 0.07786 0.07324 0.18106 0.12318
 0.2683 0.0038 0.0468 0.0493 0.0647 <.0001 0.0018
 632 640 639 638 637 640 640
  
D175 -0.0383 0.07605 0.06027 0.06024 0.05435 0.12012 -0.0148
 0.336 0.0545 0.128 0.1285 0.1707 0.0023 0.7081
 632 640 639 638 637 640 640
 
 
Pearson Correlation Coefficients 
Prob> |r| under H0:  Rho=0 
Number of Observations 
 
       
 D50 D75 D100 D125 D150 D175 
  
PLAG2 0.51392 0.49429 0.22431 0.13596 0.07867 0.06027 
 <.0001 <.0001 <.0001 0.0006 0.0468 0.128 
 639 639 639 639 639 639 
  
PLAG3 0.49127 0.47474 0.21534 0.12994 0.07786 0.06024 
 <.0001 <.0001 <.0001 0.001 0.0493 0.1285 
 638 638 638 638 638 638 
  
PLAG4 0.4703 0.4575 0.20683 0.12776 0.07324 0.05435 
 <.0001 <.0001 <.0001 0.0012 0.0647 0.1707 
 637 637 637 637 637 637 
  
D10 0.7414 0.6829 0.4271 0.34235 0.18106 0.12012 
 <.0001 <.0001 <.0001 <.0001 <.0001 0.0023 
 640 640 640 640 640 640 
  
D25 0.89453 0.82711 0.45079 0.34809 0.12318 -0.0148 
 <.0001 <.0001 <.0001 <.0001 0.0018 0.7081 
 640 640 640 640 640 640 
 
 
 
 80
 D50 D75 D100 D125 D150 D175 
      
D50 1 0.95947 0.51248 0.39095 0.03135 -0.1961 
 <.0001 <.0001 <.0001 0.4284 <.0001  
 640 640 640 640 640 640 
  
D75 0.95947 1 0.60244 0.46181 0.0617 -0.178 
 <.0001 <.0001 <.0001 0.1189 <.0001 
 640 640 640 640 640 640 
  
D100 0.51248 0.60244 1 0.84882 0.26115 0.10183 
 <.0001 <.0001  <.0001 <.0001 0.0099 
 640 640 640 640 640 640 
  
D125 0.39095 0.46181 0.84882 1 0.58413 0.36127 
 <.0001 <.0001 <.0001 <.0001 <.0001 
 640 640 640 640 640 640 
  
D150 0.03135 0.0617 0.26115 0.58413 1 0.81106 
 0.4284 0.1189 <.0001 <.0001 <.0001 
 640 640 640 640 640 640 
  
D175 -0.1961 -0.178 0.10183 0.36127 0.81106 1 
 <.0001 <.0001 0.0099 <.0001 <.0001  
 640 640 640 640 640 640 
 
 
 
 81
Appendix 4.6.1  Precipitation and temperature over the range of available data 
 
Year Precipitation (cm) 
Avg. Maximum 
Temperature (?C) 
Avg. Minimum 
Temperature (?C)  
1964 195.56 26 14 
1965 157.86 25 12 
1966 152.27 25 11 
1967 133.22 26 11 
1968 112.27 25 11 
1969 176.78 25 11 
1970 195.83 25 11 
1971 111.53 25 12 
1972 145.92 27 13 
1973 155.96 26 13 
1974 136.14 26 12 
1975 229.74 24 11 
1976 124.08 25 11 
1977 140.84 27 13 
1978 144.53 29 15 
1979 172.34 25 12 
1980 167.89 26 12 
1981 126.37 26 11 
1982 137.82 27 13 
1983 183.49 26 12 
1984 133.12 27 13 
1985 172.75 27 14 
1986 120.63 27 14 
1987 129.52 28 14 
1988 161.85 26 13 
1989 190.50 27 15 
1990 120.88 28 15 
 
 
 
 
 
 82
Appendix 5.1.1 Height growth analysis and plots 
Table 5.1.1.1  Testing overall height growth from March 1969 through December 1970 
using GLM procedure 
 
 
Source DF Sum of Squares Mean Square F Value Pr > F
  
Model 3 0.54693174 0.18231058 2.62 0.0666
  
Error 34 2.36562954 0.6957734  
    
Corrected Total 37 2.91256128    
 
R-Square Coeff Var Root MSE HGTD Mean 
0.187784 12.57756 0.263775 2.097189 
 
Source DF Type III SS Mean Square F Value Pr > F
    
SOIL 1 0.00380718 0.00380718 0.05 0.8146 
SHADE 1 0.00733278 0.00733278 0.11 0.7474 
SOIL*SHADE 1 0.54488824 0.54488824 7.83 0.0084* 
*Means significantly different (?=0.05) 
 
                            
The test shows a significant shade/soil interaction. 
 
 
 
Table 5.1.1.2  Testing normality of the residuals using the UNIVARIATE Procedure 
 
Tests for Normality 
 
Test --Statistic--- -----p Value------ 
Shapiro-Wilk W 0.953729 Pr < W 0.1182
Kolmogorov-Smirnov D 0.097890 Pr > D >0.1500
Cramer-von Mises W-Sq 0.059277 Pr > W-Sq >0.2500
Anderson-Darling A-Sq 0.453636 Pr > A-Sq >0.2500
 
Univariate tests show no significant deviations from normality.  The Anderson-Darling 
test resulted in a non significant p-value. 
 
Figure 5.1.1.1  Testing normality of the resuiduals (res) using a QQ plot 
 
 
 
 
QQ plot shows slight deviations from the line with tails, but the normality assumption is 
not violated. 
 83
 
 
Table 5.1.1.3  Testing homogeneity of variance of the residuals using Levene?s and 
Bartletts?s tests 
 
                Levene's Test for Homogeneity of HGTD Variance 
 
Source DF Sum of Squares Mean Square F Value Pr > F 
     
SOIL 1 0.000605 0.000605 0.04 0.8410 
Error 36 0.533100 0.533100   
 
Bartlett's Test for Homogeneity of HGTD Variance 
 
Source DF Chi-Square Pr > ChiSq 
 
SOIL 1 0.0425
 
0.8369 
 
 Levene?s and Bartlett?s tests for homogeneity of variance show no violations to the 
assumptions at this level of significance. 
Figure 5.1.1.2  Plot of predicted (PRED) values against residuals (RES) for                     
overall height growth 
 
This plot does show a gap between residuals, but there does not seem to be a violation of 
the homogeneity of variance. 
 84
 
 
 85
Table 5.1.1.4  Testing overall height growth from March of 1969 through  December of 
1970 using PROC MIXED 
 
The Mixed Procedure 
Type 3 Tests of Fixed Effects 
 
Effect         DF F Value Pr > F 
TIME 1069 4755.63 <.0001* 
SOIL 1069 0.24 0.6255 
SHADE 1069 0.10 0.7558 
SOIL*SHADE 1069 12.67 0.0004* 
TIME*SHADE 1069 0.10 0.7493 
TIME*SOIL 1069 0.01 0.9277 
TIME*SOIL*SHADE 1069 16.37 <.0001* 
*Means significantly different (?=0.05) 
The output from PROC MIXED shows the same significant shade/soil interaction 
as the PROC GLM output.  The normality of the residuals was tested using the same 
procedure used with the PROC GLM output. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5.1.1.3  Plot of predicted values against residual for overall height growth using 
PROC MIXED procedure 
 
 
The plot shows some systematic error, but the assumptions do not seem to be violated. 
 
                            
Figure 5.1.1.4 Example interaction plot for soil type using PROC GLM output  
 
 
 86
 
 
To test the significant interaction from the PROC GLM output, interaction plots were 
created.  This plot shows a significant interaction.  The same plot was produced for the 
shade treatment, which showed an almost identical relationship.   
 
Figure 5.1.1.5 Example interaction plot for the shade treatment using PROC GLM output  
 
                           
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 87
 
 
 88
Table 5.1.1.5   Testing the for shade/soil interactions using LSD (least significant 
difference) multiple comparison 
 
 
t-tests (LSD) for HGM 
NOTE: This test controls the Type I comparisonwise error rate, not the 
experimentwise error rate. 
                     
Alpha 0.05 
Error Degrees of Freedom 34 
Error Mean Square 0.069577
Critical Value of t 2.03224 
Least Significant 
Difference 0.2471 
Harmonic Mean of Cell 
Sizes 9.411765
 
 
NOTE: Cell sizes are not equal. 
 
t Grouping Mean N SSC 
    
A 2.2138 10 NW 
A    
A 2.206 10 SL 
A    
B         A 1.9933 10 NL 
B    
B 1.9453 8 SW
 
 Means with the same letter are not significantly different (?=0.05) 
SSC = shade/soil combination, NL = no shade, Lucy soil; NW = no shade, Wagram soil; 
SL = shade, Lucy soil; SW = shade, Wagram soil 
 
 
This test shows a significant difference in height growth for the shade treatment on 
Wagram soils.  There is also a significant difference in height growth for shaded trees on 
both soil types. 
 
 
 
 
 89
Table 5.1.1.6   Testing the for shade/soil interactions using a ttest to test for significant 
differences between shade treatments within the Wagram soil type 
       
  
       
   
Lower 
CL  Upper CL 
Lower 
CL 
Variable SHADE N Mean Mean Mean Std Dev 
       
HGM N 10 2.0903 2.2138 2.3373 0.1188
HGM S 8 1.6781 1.9453 2.2124 0.2113
HGM Diff (1-2)  0.0192 0.2685 0.5178 0.1846
      
       
  Upper CL     
Variable SHADE Std Dev Std Err Minimum Maximum  
      
HGM N 0.3153 0.0546 1.997 2.556  
HGM S 0.6504 0.1130 1.385 2.324  
HGM Diff (1-2) 0.3773 0.1176    
    
   
                                    T-Tests     
   
Variable Method Variances DF t Value Pr > |t|  
      
HGM Pooled Equal 16 2.28 0.0364*  
HGM Satterthwaite Unequal 10.2 2.14 0.0575  
      
 
                                        Equality of Variances    
    
Variable Method Num DF Den DF F Value Pr > F  
      
HGM Folded F 7 9 3.42 0.0899  
 
Means with an * are significantly different (?=0.05) 
 
 
This test reinforces the significant the LSD test for the significant interaction. 
 
 
 
 
 
Figure 5.1.1.6 Plot of overall height growth means by shade/soil combination with two 
standard deviations:  (SSC= shade/soil combination, NL = no shade, Lucy soil;  
NW = no shade, Wagram soil;SL = shade, Lucy soil; SW = shade, Wagram soil) 
 
 
 
 
 
 
 90
 
 
Appendix 5.4.1 Diameter growth plots 
Diameter growth was evaluated over each year separately because of calibration 
issue during a January 1970 measurement and because of negative growth values.  All 
tests and assumption diagnostics were conducted using the same procedures outlined in 
Appendix 5.1.1. 
Figure 5.4.1.1  Average monthly diameter measurements by soil type 
 
3.
54
4.
55
5.
56
6.
5
Ma
r
-
69
Ap
r
-
69
Ma
y
-
69
J
un- 69
Ju
l
-
69
A
ug-
69
S
ep-
69
Oc
t
-
69
No
v
-
69
De
c
-
69
Ja
n
-
70
F
eb-
70
Ma
r
-
70
Ap
r
-
70
Ma
y
-
70
Ju
n
-
70
Ju
l
-
70
A
ug- 70
Se
p
-
70
Oc
t
-
70
No
v
-
70
De
c
-
70
M
o
nt
hs
A ve r ag e D i am et e r s ( cm )
Lu
c
y
 l
o
am
y
 
s
an
d
W
a
gr
a
m
 l
o
am
y
 s
an
d
 91
 
 
Figure 5.4.1.2   Average diameter growth by shade/soil combination for 1969 showing 
two standard errors: (SSC= shade/soil combination, NL = no shade, Lucy soil;  
NW = no shade, Wagram soil;SL = shade, Lucy soil; SW = shade, Wagram soil) 
 
 
 
 
 
 
 92
 
 
 
Figure 5.4.1.3   Average diameter growth by shade/soil combination for 1970 showing 
two standard errors: (SSC= shade/soil combination, NL = no shade, Lucy soil;  
NW = no shade, Wagram soil;SL = shade, Lucy soil; SW = shade, Wagram soil) 
 
 
 
 93
 
 
 94
Appendix 5.6.1 Diurnal and nocturnal growth example regression 
This is an example regression procedure.   
 
Table 5.6.1.1 Regression model for dependent variable: mean nocturnal growth rate and 
independent variable:  D55T 
 
Analysis of Variance 
 
Source DF Sum of Squares Mean Square F Value Pr > F 
      
Model 1 0.00131 0.00131 21.84 0.0002
Error 17 0.00102 0.00006019  
Corrected Total 18 0.00234  
 
Root MSE 0.00776 R-Square 0.5623 
Dependent Mean 0.04642 Adj R-Sq 0.5365 
Coeff Var 16.71326  
 
 
Parameter Estimates 
 
Variable  DF 
Parameter 
Estimate 
Standard 
Error t Value Pr > | t |
     
Intercept  1 0.01638 0.00667 2.46 0.0251 
D55T  1 0.00010628 0.00002275 4.67 0.0002 
 
 
 
 
 
 
 
 
 
 
Table 5.6.1.2  UNIVARIATE Procedure tests for rormality 
 
Test --Statistic--- -----p Value------ 
Shapiro-Wilk W 0.969524 Pr < W 0.7669
Kolmogorov-Smirnov D 0.160899 Pr > D >0.1500
Cramer-von Mises W-Sq 0.047908 Pr > W-Sq 0.2500
Anderson-Darling A-Sq 0.261653 Pr > A-Sq 0.2500
 
Univariate tests show not significant deviations from normality.  The Anderson-Darling 
test resulted in a non significant p-value. 
Figure 5.6.1.1  QQ plot 
 
QQ plot shows slight deviations from the line with tails, but the normality assumption is 
not violated. 
 
 
 
 
 
 
 95
 
 
Figure 5.6.1.2 Plot of predicted values against residual for regression model 
Dependent Variable: Mean Nocturnal Growth Rate 
Independent Variable:  D55T 
 
 
 
 
 
This plot does not show a violation of the homogeneity of variance because of the even 
scatter of residuals. 
 
 
 96
 
 
Figure 5.6.1.3  Plot of  regression line 
Dependent Variable: Mean Nocturnal Growth Rate 
Independent Variable:  D55T 
 
 
 
The plot shows even scatter of points along the regression line with no significant 
deviations from the assumptions.
 97
 
 
 98
Appendix  5.7.1  Yearly measurements analysis and plots 
Table 5.7.1.1  Checking soil condition using nonparametric tests 
Wilcoxon Scores (Rank Sums) for Variable Height Growth 1969 through 1981 
Classified by Variable SOIL 
 
SOIL N 
Sum of 
Scores 
Expected  
Under HO 
Std Dev.  
Under HO Mean Score 
L 17 336 306 30.298515 19.764706 
W 18 294 324 30.298515 16.333333 
 
                     
Wilcoxon Two-Sample Test 
 
Statistic 3.36E+02
 
Normal Approximation 0.9736
One-Sided Pr > Z 0.1651
Two-Sided Pr > |Z| 0.3302
  
t Approximation  
One-Sided Pr > Z 0.1686
Two-Sided Pr > |Z| 0.3371
 
Z includes a continuity correction of 
0.5  
 
 
Kruskal-Wallis Test  
Chi-Square 0.9804
DF 1
Pr > Chi-Square 0.3221
 
The test showed no significant difference in mean height growth by soil type because the 
pvalues (Pr >) were greater than the 0.05 level of significance.
 
 
 99
Figure 5.7.1.1  Average yearly height measurements from 1969 through 1981 
 
 
 
Figure 5.7.1.2  Average yearly height growth from 1969 through 1981 
 
 
 100
 
 
Figure 5.7.1.3  Average yearly diameter growth from 1972 through 1981 
 
 
 
 101
 
 
Figure 5.7.1.4  Average yearly diameter growth from 1972 through 1981 
 
                 
 102