Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and testing of a weatherometer to accelerate the surface checking of wood Ratu, Ricky Novry 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_spring_ratu_ricky.pdf [ 6.98MB ]
Metadata
JSON: 24-1.0067135.json
JSON-LD: 24-1.0067135-ld.json
RDF/XML (Pretty): 24-1.0067135-rdf.xml
RDF/JSON: 24-1.0067135-rdf.json
Turtle: 24-1.0067135-turtle.txt
N-Triples: 24-1.0067135-rdf-ntriples.txt
Original Record: 24-1.0067135-source.json
Full Text
24-1.0067135-fulltext.txt
Citation
24-1.0067135.ris

Full Text

DEVELOPMENT AND TESTING OF A WEATHEROMETER TO ACCELERATE THE SURFACE CHECKING OF WOOD  by  RICKY NOVRY RATU B. Eng. (S.T.)., Universitas Sam Ratulangi, 1993  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (FORESTRY)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2009 © Ricky Novry Ratu, 2009  ABSTRACT A new version of a weathering device (Accelerated Check Tester, ACT) was built, and weathering cycles for accelerating the surface checking of decking board samples exposed in the device were developed. The device permits the testing of realistic-sized decking board samples that are oriented horizontally and restrained in the device by screws. Two experiments were carried out to validate the device and associated test cycles. In the first experiment, southern pine (Pinus sp.) and western red cedar (Thuja plicata) samples were exposed to 6 different cycles in the ACT to determine which elements of weathering cycles (moisture, heat, freezing and UV radiation) were critical to the ACT’s function of accelerating checking. Large number of checks developed on the surface of samples subjected to wetting and drying cycles. Samples subjected to a cycle that also involved UV exposure developed significantly more and larger checks than samples subjected to any of the others cycles. Checking was much more pronounced in southern pine than in western red cedar samples. The second experiment examined the checking and distortion that developed in artificially weathered boards compared to those that developed in naturally weathered boards. Matched pairs of kerfed and unkerfed southern pine boards were subjected to accelerated weathering in the ACT or natural weathering. The number and length of checks that developed in boards exposed in the ACT were similar but not identical to those in boards subjected to natural weathering. The width of checks was greater in boards subjected to accelerated weathering. Kerfing had no significant effect on checking or distortion of boards. Exposure of boards in the ACT accelerates checking approximately 16 times compared to natural weathering. The Accelerated Check Tester should be a very useful tool for obtaining information on factors that affect the checking of wood.  ii  TABLE OF CONTENTS Abstract ........................................................................................................................... ii Table of Contents ........................................................................................................... iii List of Tables ................................................................................................................. vi List of Figures ............................................................................................................... vii Acknowledgements ......................................................................................................... x Chapter One GENERAL INTRODUCTION ....................................................................................... 1 1.1  Introductory Remarks .......................................................................................... 1  1.2  Hypothesis............................................................................................................ 3  1.3  Aim and Importance ............................................................................................ 3  1.4  Study Outline ....................................................................................................... 4  Chapter Two SURFACE CHECKING OF WOOD ............................................................................ 5 2.1  Theory .................................................................................................................. 5  2.2  Factors Affecting Checking ................................................................................. 8  2.2.1  Environmental Factors ................................................................................... 8  2.2.1.1 2.2.1.2 2.2.1.3 2.2.1.4  2.2.2 2.2.3  Light .................................................................................................................... 9 Water ................................................................................................................. 11 Heat ................................................................................................................... 12 Other Factors..................................................................................................... 12  Wood Species ............................................................................................... 13 Wood Properties ........................................................................................... 15  2.3 Treatments to Reduce Checking ........................................................................ 17 2.3.1 Water Repellents .......................................................................................... 17 2.3.2 Hydrophobic Wood Preservatives ................................................................ 19 2.3.3 Mechanical Treatments ................................................................................ 22 iii  2.3.3.1  Incising.............................................................................................................. 22  2.3.3.2  Through Boring and Center Boring .................................................................. 23  2.3.3.3  Kerfing .............................................................................................................. 25  2.4 Accelerated Weathering and Checking .............................................................. 27 2.4.1 Introduction .................................................................................................. 27 2.4.2 Types of Accelerated Weathering Tests / Machines .................................... 28 2.4.3 Correlation between Natural and Artificial Weathering .............................. 30 2.4.4 Accelerated Weathering Devices and Cycles to Increase Checking of Wood ............................................................................................................ 31 2.5  Summary ............................................................................................................ 34  Chapter Three CYCLE OPTIMIZATION TO ACCELERATE SURFACE CHECKING OF DECKING BOARDS IN A CUSTOMIZED WEATHERING DEVICE ................... 36 3.1  Introduction ........................................................................................................ 36  3.2  Development of the Accelerated Check Tester ................................................. 37  3.2.1 3.2.2 3.2.3  Hardware Used to Construct the Accelerated Check Tester ........................ 37 Control System ............................................................................................. 40 Ancillary Equipment .................................................................................... 41  3.3 Optimization of Weathering Cycle .................................................................... 42 3.3.1 Experimental Design .................................................................................... 42 3.3.2 Preparation of Decking Boards .................................................................... 43 3.3.3  Accelerated Weathering ............................................................................... 45  3.3.3.1  3.3.4  Cycle ................................................................................................................. 45  Measurement of Checking and Distortion.................................................... 48  3.4 Results ................................................................................................................ 51 3.4.1 Effect of Cycle Type on Checking of Boards .............................................. 51 3.4.2 Effect of Wood Species on Checking of Boards .......................................... 55 3.4.3 Effect of Cycle Type on Shape Distortion of Boards ................................... 57 3.4.4 Effect of Wood Species on Shape Distortion of Boards .............................. 59  iv  3.5  Discussion .......................................................................................................... 62  3.6  Conclusions ........................................................................................................ 64  Chapter Four COMPARISON OF CHECKING IN SOUTHERN PINE SAMPLES SUBJECTED TO ACCELERATED WEATHERING AND NATURAL EXPOSURE .................... 65 4.1  Introduction ........................................................................................................ 65  4.2  Experimental Design .......................................................................................... 66  4.3 Materials and Methods ....................................................................................... 68 4.3.1 Sample Preparation....................................................................................... 68 4.3.2 Decking Racks .............................................................................................. 70 4.3.3 Weathering Schedules .................................................................................. 70 4.3.3.1  Natural Weathering ........................................................................................... 70  4.3.3.2  Accelerated Weathering .................................................................................... 72  4.4 Results ................................................................................................................ 74 4.4.1 Checking ....................................................................................................... 74 4.4.2 Distortion ...................................................................................................... 81 4.5  Discussion .......................................................................................................... 84  4.6  Conclusions ........................................................................................................ 86  Chapter Five GENERAL CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH ................................................................................................................. 88 5.1 5.2  General Conclusions .......................................................................................... 88 Suggestions for Further Research ...................................................................... 91  References .................................................................................................................... 92 Appendices ................................................................................................................. 104 Appendix 1 .............................................................................................................. 104 Appendix 2 .............................................................................................................. 118 v  LIST OF TABLES  Table 2.1: Location of checks in selected softwoods subjected to artificial weathering (Source: Coupe and Watson 1967) .............................................. 14 Table 2.2: Summary of previous studies that have employed accelerated devices to examine the weathering of wood ................................................................. 33 Table 3.1: Density and growth ring/cm of sample boards ................................................ 44 Table 3.2: Combination of weathering factors used in the different weathering cycles .................................................................................................................... 45 Table 3.3: Steps involved in the six different accelerated weathering cycles ................ 46 Table 3.4: Significant effects of, and interactions between, weathering cycle and wood species on checking and distortion of samples exposed in the Accelerated Check Tester to different weathering cycles ............................. 51 Table 4.1: Wood characteristics of southern pine samples ............................................... 69 Table 4.2: Steps involved in the artificial weathering cycle............................................. 72 Table 4.3: Average check length of deck board samples subjected to artificial weathering............................................................................................................ 80 Table 4.4: Average check length of deck board samples subjected to natural weathering............................................................................................................ 80  vi  LIST OF FIGURES  Figure 2.1: Mechanism of check formation in a restrained decking board (Source: Evans 2008) ................................................................................. 7 Figure 2.2: Orientation of growth rings to the upper exposed surface of decking boards; convex orientation, left; concave orientation, right ...................... 17 Figure 2.3: Kerfing and center-boring treatments (Evans et al. 2000) ........................... 24 Figure 3.1: Technical drawing of the Accelerated Check Tester.................................... 37 Figure 3.2: Three main parts of the Accelerated Check Tester....................................... 38 Figure 3.3: KWIKI decking screws used for fixing specimens to hardwood supports in the Accelerated Check Tester .................................................... 39 Figure 3.4: Image of the ACT showing the water spraying system ............................... 40 Figure 3.5: UV curing machine used to irradiate samples subjected to cycle 6 ........... 42 Figure 3.6: Dimensions of southern pine and western red cedar decking samples subjected to accelerated weathering .............................................................. 44 Figure 3.7: Cyclical sequences of steps during accelerated weathering ........................ 47 Figure 3.8: Tools used for measuring the length and width of checks........................... 49 Figure 3.9: Digital gauge and aluminium table used for assessing shape distortion of boards ........................................................................................................... 50 Figure 3.10: The effect of cycle type on the total number of checks (averaged across southern pine and western red cedar samples) ................................ 52 Figure 3.11: The effect of cycle type on total check width (averaged across southern pine and western red cedar samples) ............................................ 52 Figure 3.12: The effect of cycle type on total check length (averaged across southern pine and western red cedar samples) ............................................ 53 Figure 3.13: The effect of cycle type on check length/width ratio (averaged across southern pine and western red cedar samples) ............................................ 53 Figure 3.14: Scanned images of southern pine samples after exposure to different weathering cycles ............................................................................................ 54 Figure 3.15: Scanned images of western red cedar samples after exposure to different weathering cycles ............................................................................ 54  vii  Figure 3.16: The effect of wood species on check number (averaged across cycle type) ................................................................................................................... 55 Figure 3.17: The effect of wood species on total check width (averaged across cycle type) ........................................................................................................ 56 Figure 3.18: The effect of wood species on total check length ........................................ 56 Figure 3.19: The effect of wood species on check length/width ratio (shape)............... 57 Figure 3.20: The effect of cycle type on cupping (averaged across southern pine and western red cedar samples) ..................................................................... 58 Figure 3.21: The effect of cycle type on bow (averaged across southern pine and western red cedar samples) ............................................................................ 58 Figure 3.22: The effect of cycle type on twist (averaged across southern pine and western red cedar samples) ............................................................................ 59 Figure 3.23: The effect of wood species on cupping (averaged across cycle type) ...... 60 Figure 3.24: The effect of wood species on bow (averaged across cycle type) ............. 60 Figure 3.25: The effect of wood species on twist (averaged across cycle type) ............ 61 Figure 3.26: Shape distortion of western red cedar (left) and southern pine (right) samples after weathering (Cycle 4) ............................................................... 61 Figure 4.1: Allocation of specimens to experimental factors ........................................... 68 Figure 4.2: Kerfed and unkerfed samples prior to weathering ......................................... 69 Figure 4.3: Kerfed and unkerfed decking board specimen (arrowed) fastened to a decking rack ..................................................................................................... 70 Figure 4.4: Decking board specimens exposed to natural weathering ............................ 71 Figure 4.5: The total number of checks in specimens subjected to artificial or natural weathering ........................................................................................... 74 Figure 4.6: The total length of checks in specimens subjected to artificial or natural weathering ........................................................................................................ 75 Figure 4.7: The total width of checks in specimens subjected to artificial or natural weathering ........................................................................................................ 76 Figure 4.8: The average check length of specimens subjected to artificial or natural weathering ........................................................................................................ 77 Figure 4.9: Distribution of check widths in specimens subjected to artificial or natural weathering ........................................................................................... 78 Figure 4.10: Images of kerfed and unkerfed boards (block 5) after 5 days of artificial weathering or 26 weeks of natural weathering ............................ 78  viii  Figure 4.11: Check number in specimens exposed to natural weathering compared to the number (dashed line) that developed in specimens exposed to 5 days of artificial weathering........................................................................... 79 Figure 4.12: Total length of checks in specimens exposed to natural weathering compared to the total length of checks (dashed line) that developed in specimens exposed to5 days of artificial weathering ................................. 79 Figure 4.13: The effect of weathering method on cupping of boards ............................. 82 Figure 4.14: The effect of weathering method on bowing of boards .............................. 82 Figure 4.15: The effect of weathering method on twisting of boards ............................. 83  ix  ACKNOWLEDGEMENTS I would like to thank all people who helped and inspired me during my study at the Department of Wood Science, Faculty of Forestry the University of British Columbia. I especially want to thank my supervisor, Dr. Philip David Evans, for the time he dedicated to me and for his guidance during my research. It has most certainly been a fruitful journey since the 38th IRG meeting in Jackson, Wyoming USA. I sincerely thank him for providing me with an opportunity to complete a Master’s degree. His perpetual energy, enthusiasm for research and his expertise motivated me to discover a new field study that I never thought of before. Dr. Stavros Avramidis and Dr. Taraneh Sowlati deserve a special thanks as my thesis committee members. Their advice in the committee meeting and also their encouragements in informal meeting including the weekly social activities (donut’s Thursday) were a great help. I was delighted to interact with Dr. Susan B. Watts by attending her classes, and having her on my examination committee. I will always remember the exciting field trip to the interior of BC during the FRST 547 course that broadened my knowledge about forestry. I am grateful to all members of Dr Evans’ research group for making the lab a convivial place to work. The generous support from Dr. Mohammed J. A. Chowdhury in all aspects of my academic life and his generosity to my family is greatly appreciated. I would like to thank Ian Cullis for his friendship and his assistance with all types of technical problems. Also thanks to Dr. Kate Semple and Dr. Marcos Gonzales for their advice. Thanks to my former fellow graduate students, Guenter Modzel and Martin Miesner, for their camaraderie during my research. My thanks also to Arash Jamali, Vicente Hernandez, and the new lab members Jonathan Haase, Stephan Vollmer, and Chunling Liu. I will always proud to be a member of this superb group that includes people from 11 different countries. Administrative staff in CAWP/Wood Science Department (UBC) and technicians in Wood machines lab have my thanks for supporting my academic research. Also thanks to all friends in the Department of Wood Science for making my research and social life fruitful. My deepest gratitude goes to my wife, Mitsi Singal, for her unflagging love and support throughout my life; and also to my kids Leri, Valdi and Verrel who will always be my inspiration. I would also like to extend my deep and sincere gratitude to my father John Ratu for his prayers and support during my overseas study. The most overall, Praise be to God forever and ever, Thy give me a strength and knowledge. He has made everything beautiful in its time. "There is surely a future hope for you, and your hope will not be cut off (Proverbs 23:18)"  x  1  Chapter one  GENERAL INTRODUCTION 1.1 Introductory Remarks Wood is the material of choice for many applications and its use depends on properties, such as strength and stiffness and its cost compared to other materials. The use of wood in some applications has increased over time. Wood decks, for example, have become an important part of residential construction in recent years and the demand for decking timber is predicted to increase over the next five years (Freedonia Group Inc. 2005). In 2000 the number of decks, porches, verandas, and balconies that were built, repaired, or replaced in the USA was estimated to be 3.2 million units, which generated a market for decking materials of 5.3 billion board feet (12.51 million m3) (Principia Partners 2002). Freedonia Group Inc. (2005) predict that US demand for decking timber will increase by more than two percent per year through 2009 to 5.6 billion board feet (13.22 million m3), valued at $5.9 billion. Furthermore, the decking market is stable, which stems from the fact that more than 85 percent of demand is generated through repair and improvement activity, which is inherently less cyclical than the new construction market (Freedonia Group Inc. 2005). The main reason why wood has captured the largest share of the decking market is because of its relatively low cost, good structural properties and durability (when treated). It is also visually appealing, environmentally friendly, easy to recycle, naturally renewable and easily processed with low energy consumption (Imamura 2001). However, a problem with the use of wood outdoors is that it weathers (Feist, 1983). Weathering is the surface degradation of wood caused by the combined effects of solar radiation in sunlight, water, heat, oxygen, environmental pollutants and abrasion by wind-blown sand or other particulates. During weathering wood surfaces lose their natural colour and become rough as grain raises and the wood checks (Feist, 1983). 1  The surface checking of wood is one of the most obvious and serious deleterious effects of weathering. Surface checks and severe cracks still develop in treated wood exposed to the weather. The checking of decking timber is disliked by customers. Accordingly, a study by McQueen and Stevens (1998) and Alderman et al. (2003) found that “aesthetics” was one of the most important factors influencing the removal of decks by homeowners. The average age of decks when they were removed from service in the USA was between 9 and 12.8 years (McQueen and Stevens 1998, Alderman et al. 2003). Other factors that contributed to the replacement of decks were the amount of decayed wood, physical degradation of the wood, and the structural integrity of the deck (Alderman et al. 2003). Therefore, the development of methods to maintain the aesthetic appeal of wood by reducing surface checking will help prolong its lifespan as a decking material and lead to improved customer satisfaction, and help to conserve wood and forest resources. There has been significant interest in developing preservatives that are better at preventing wood from checking (Jin et al. 1991, Cui and Zahora 2000, Christy et al. 2005). Currently, however, there is no accepted test methodology for assessing the checking of treated timber or accelerating the development of checks so information on the effectiveness of treatments at restricting checking can be obtained more quickly. Methods for accelerating the weathering of materials rely on the use of devices (weatherometers) that expose the material to factors that cause surface weathering (Arnold et al. 1991, Suits and Hsuan 2003). Weatherometers have been widely used to evaluate the ability of chemical treatments to restrict the weathering of wood (Rowell et al. 1981, Feist and Williams 1991), but they cannot easily accommodate large dimension samples such as decking boards, and they do not permit horizontal orientation of samples. Furthermore, most of the weathering cycles that have been developed for weatherometers seek to maximize the surface degradation of materials rather than generate anisotropic distribution of stress in samples that causes the checking of wood. Consequently, a different approach is needed to accelerate surface checking of wood.  2  1.2 Hypothesis Most researchers attribute surface checking of wood to the development of surface stresses and strains generated by shrinkage and swelling due to moisture loss and gain within the hygroscopic range that exceeds the elastic limit of wood (Stamm 1965, Evans 2004). The exposed wood swells and shrinks and steep moisture gradients between surface and sub-surface layers of boards generate stresses that cause surface checking (Stamm 1965, Evans et al. 2008). Sandberg (1999) stated that the greater occurrence of checking on tangential surfaces is mainly the result of mechanical degradation induced by stresses, which build up due to “anisotropic moisture movement of the wood material and moisture gradients between the surface and the core of softwood members”. When wood surfaces dry, the wetter inner zone shows less tendency to shrink compared to the dryer surface, and stresses and strains are generated at the surface due to large moisture gradients. When these strains exceed the tensile strength/elastic limit of wood perpendicular to the grain, checking of the wood surface occurs (Stamm 1965). Therefore, it seems reasonable to hypothesize that checking is related to severity of moisture gradients and differences in shrinkage and swelling between surface and core. Therefore, a device that can artificially induce such changes and their frequency of occurrence should increase the surface checking of wooden decking boards.  1.3 Aim and Importance The aim of this study was to complete the development of a device, including designing optimal weathering cycles, to accelerate the surface checking of decking timber. This overall aim was achieved by: (1) Developing a new version of a device (Accelerated Check Tester, ACT) designed to accelerate the checking of wooden decking samples; (2) Determining which elements of weathering cycles (heat, wetting and drying, UV radiation, freezing) influence surface checking; (3) Comparing the  3  checking that developed in samples exposed to accelerated weathering in the ACT with that occurring in matched samples exposed to natural weathering. This study is important because the development of a device and associated weathering cycles that can accelerate the surface checking of wooden decking boards could reduce the time required to test new wood preservatives and finishes designed to prevent the checking of decking boards. Reduction in the time required to develop new timber treatments not only lowers the cost of developing the treatment, which is important to industry, but also allows greater numbers of treatments to be tested. This increases the probability of developing treatments that are more effective at preventing the checking of decking boards. Such treatments are needed because checking of wood is disliked by consumers and this has allowed lumber substitutes such as wood plastic decking to capture an increasing share of the market for decking products.  1.4 Study Outline This Chapter (Chapter 1) provides the rationale for the study. Chapter 2 reviews the relevant literature on the surface checking of wood, including importance of checking, factors affecting checking such as environment, wood species, and wood properties. Chapter 2 also reviews the chemical and mechanical treatment such as incising, kerfing and center boring that have been used to reduce the severity of checking of wood products. The use of accelerated weathering devices and cycles to increase the surface checking of wood are also reviewed. In Chapter 3, the development of a device and weathering cycles to accelerate surface checking of wood is described. Chapter 4 describes an experiment that was carried out to compare checking of samples in the accelerated weathering device with that occurring during natural weathering. Finally, Chapter 5 discusses results from the two experimental chapters, draws conclusions and makes suggestions for further research.  4  2 2 Chapter Two SURFACE CHECKING OF WOOD  One of the main weaknesses of wood used outdoors is its sensitivity to environmental degradation. When wood is exposed outdoors and above ground, a complex decomposition process occurs at the surface of the material due to chemical, physical, biological and mechanical factors (Sandberg 1999).  This is a complex  process involving the interaction of several environmental factors such as light and moisture, and is termed weathering (Feist 1990). This chapter reviews the existing scientific information on the weathering of wood with particular emphasis on checking of wood and the methodology for accelerating the weathering and checking of wood.  2.1 Theory Flæte et al. (2000) define checks as cracks, which do not penetrate wood deeper than 75 % of the board thickness. Checks can be classified as macrochecks, which are visible to the naked eye, and microchecks, which can only be observed under the microscope. Checks are three-dimensional structures, and can be characterized by their length, width, depth, volume and specific shape. Checking is disliked by end users of wood, because it detracts from the appearance of wood and it also contributes to roughening of wood surfaces, and may increase the susceptibility of wood to biological deterioration. Checking is closely related to physical degradation and loss of woods aesthetic appeal (Feist and Mraz 1978). Hence, severe checking may result in replacement of wood used outdoors. Generally, checking is caused by differential rates of swelling and shrinkage in surface and sub-surface zones within wood (Schniewind 1963, Stamm 1965, Evans 2008). Swelling and shrinking occur as the moisture content of wood changes in  5  response to daily as well as seasonal changes in the relative humidity and temperature of the surrounding atmosphere (Eckelman 1998). The shrinkage and swelling of wood are proportional to the volume of water lost and gained up to a level known as the fiber saturation point. Swelling and shrinkage both create stresses within wood. The stresses and strains that lead to the checking of wood occur as a result of shrinkage anisotropy, and in particular the difference between radial and tangential shrinkage of wood (McIntosh 1955). In untreated timber, moisture and stress gradients are generated from the surface to the core of the wood and cracks may be formed as deep as 10 mm below the surface. In addition the increased moisture content at the surface weakens the wood, which leads to additional checking (Bariska et al. 1988). During weathering, wood is subjected to wetting and drying and most of the checks, which develop during initial drying, reopen and in some cases increase in length (Mackay 1973). The following section describes how checks develop at the surface of wood exposed outdoors. Wood exposed outdoors adjusts its moisture content to reach an equilibrium with its surrounding atmosphere. This moisture content is referred to as the equilibrium moisture content of wood (Eckelman 1998). It follows that for different temperatures and levels of relative humidity of the air, there is a corresponding equilibrium moisture content of wood (Eckelman 1998). The outer layer is affected first by changes in relative humidity whereas the underlying core is less affected. This observation is very relevant to wooden boards in decks exposed outdoors. Since decks are usually exposed horizontally to the weather, their top surface is directly affected by water and solar radiation, whereas the underlying wood is protected to varying degrees (Urban and Evans 2005, Weizenegger 2006, Evans et al. 2008).  6  Figure 2.1: Mechanism of check formation in a restrained decking board (Source: Evans 2008) Figure 2.1 illustrates the process of checking of wooden decking boards exposed outdoors. The upper side of boards is most affected by rain and dew as mentioned above. The water eventually penetrates into the core of the board (Figure 2.1, 1). Solar radiation heats the surface of boards; temperatures up to 80 °C have been measured (Wengert 1966, Sell and Walchli 1969, Weizenegger 2006, Evans 2008) and water is removed from the surface of the board (Figure 2.1, 2). As a result the moisture content of the surface drops below the fiber saturation point and the surface starts to shrink (Figure 2.1, 3). The core of the board, however, stays in a green condition for a longer time and restrains the surface from shrinking (Evans 2008). In addition, fasteners restrain boards and keep them from cupping and warping (Figure 2.1, 4). The shell of boards is put under tension (blue arrow), while the core is under compression (red arrow). A state may be reached where the tension stress at the surface exceeds the tensile strength of the surface cells perpendicular to the grain and then the wood checks (Stamm 1965, Sell and Walchi 1969, Evans 2008) (Figure 2.1.5). 7  2.2 Factors Affecting Checking During exterior exposure checking often occurs at wood surfaces and such checks are one of the first visible signs of the physical deterioration of wood (Feist and Mraz 1978). The causes of checking are manifold. Most previous studies suggest that environmental factors such as solar radiation, water, and heat that cause the weathering of wood also influence checking. Nevertheless, the factors involved in weathering act in different ways. Moreover, solar radiation and moisture sometimes affect the exposed wood at different times (Kühne et al. 1968, Borgin 1971). Thus, degradation pathways may vary depending on the interplay of the elements, with solar radiation enhancing the effect of water or the converse (Feist 1990). Other factors that influence the checking of wood are wood species and their properties. In addition, some studies have shown that checks may occur prior to drying or even develop in wood during harvesting and machining (Schniewind 1963). 2.2.1  Environmental Factors The susceptibility of wood to weathering is generally seen as one of the  greatest drawbacks to its use outdoors (Coupe and Watson 1967, Sell and Leukens 1969, Kamden and Zhang 2000). The grey colouration and surface roughness of weathered wood are its most obvious features (Sell 1968, Chang et al. 1982, Feist 1982, 1983, Jin et al. 1991, Evans 2001). As pointed out by Feist and Hon 1984 “the roughened surface of exposed wood eventually gathers dirt and mildew. Surface checks grow into larger cracks. Grain may loosen and boards eventually cup and warp and pull away from fasteners, and the wood may become unsightly.” Groves and Banana (1986) mentioned that it is difficult to discriminate between the effects of the three agencies involved in weathering; sunlight, water and oxygen. The exposure of wood to water and light together results in a surface appearance, which differs from that caused by light or water alone (Feist 1982).  8  2.2.1.1 Light Solar radiation consists of ultraviolet (200-380 nm), visible (380-700 nm) and infrared (700-2200 nm) radiation. Wood is an excellent light absorber. It absorbs light strongly below 200 nm, has some absorption maxima between 200 and 300 nm and a tail of absorption extending to 400 nm and then into the visible region of the spectrum (Hon 1981). Wood is sensitive to light (Hon and Chang 1984) and most of its components that absorb visible and ultraviolet light are susceptible to photochemical degradation. This leads to discoloration, loss of gloss and brightness, and the production of various degradation products. Even a short period of exposure to sunlight leads to pronounced losses of surface integrity (Miller and Derbyshire 1981). The effect of light on wood occurs over a wide range between the ultraviolet (UV) and the infrared (IR) spectrum (Miller and Derbyshire 1981). UV light contains the most destructive wavelengths and has the strongest effect on the discolouration and degradation of wood (Browne and Simonson 1957, Derbyshire and Miller 1981). Owen et al. (1993) concluded that the effect of UV light on wood surfaces is quite rapid; noticeable effects are seen after 50 h of light-only artificial weathering. They observed that the rate of degradation of wood exposed to light is approximately 20 times greater in air than in an atmosphere of pure nitrogen. UV light is responsible for the primary photooxidative degradation of wood even though it only accounts for 5% of the total solar energy received at the earth’s surface (Miller and Derbyshire 1981, Hon 1983, Feist and Hon 1984). Solar radiation affects the surface layers of wood because light is only capable of penetrating wood to a depth of 0.05-2 mm (Feist and Hon 1984). Anderson et al. (1991) mentioned that the penetration of UV light into wood is restricted to about 75 m. Nevertheles, Kataoka and co-workers (2004) found that after 1500 hours exposure to artificial sunlight, photochemical changes in sugi (Cryptomeria japonica. D. Don) earlywood occurred to a depth of 700  m.  Light degrades lignin, cellulose and  hemicellulose via free radical mechanisms, which results in deterioration of the wood cell wall (Derbyshire and Miller 1981; Hon et al. 1985). There is evidence that longwave (UV-A) and shortwave (UV-B) UV radiation cause the same anatomical  9  changes at exposed wood surfaces, but the rate of change occurs much faster during exposure to shortwave (UV-B) radiation (Miniutti 1967). Visible light is also capable of degrading wood. Exposure of wood to the visible component of the solar spectrum results in changes to wood’s ultrastructure, which are associated with tensile strength losses (Derbyshire and Miller 1981). Timber exposed only to visible and infrared light degrades at about half the rate observed for wood exposed to the full solar spectrum (Derbyshire and Miller 1981). Visible and infrared light causes browning of wood (Browne and Simonson 1957). Later the surface gets more and more discoloured and visible light is selectively absorbed (Hon 1981). The rate of degradation is increased further as the brown color of wood deepens, which leads to increases in the absorption of infra-red radiation (Hon 1981) and raises the rate of photochemical and oxidative reactions (Rowell et al. 1981). Light is also capable of altering wood colour by degrading extractives (Derbyshire and Miller 1981; Hon et al. 1985). Two types of mechanical degradation can be distinguished at exposed wood surfaces; the erosion and checking of single cell walls and the disintegration of the cell structure, which becomes more pronounced with increasing exposure to light (Sell and Leukens 1971a). Hon (1981) stated that photooxidation of wood surfaces causes extensive surface modification such as roughening and checking. Arndt and Willeitner (1969) attributed the stronger roughening of the surface of samples exposed high up in mountains to the increased intensity of solar radiation at high altitude. Sandberg (1999) stated that photochemical reactions at wood surfaces accelerated the propagation of checks, which were caused by stress variation in wood due to changes in moisture content. Sell and Leukens (1969) stated that checking of wood was facilitated by the photolytic weakening of the surface layers (Sell and Leukens 1969). Stamm and co-workers noted that face checking was more severe in heat-stabilized wood because prolonged exposure to infrared radiation embrittles the wood surface (Stamm 1964, 1965). A recent study of the effect of solar radiation on the development of checks at wood surfaces found that exposure of wood to UV-A, UV-B and visible light  10  increased the size and number of checks that developed (Evans et al. 2008). They found that decking samples exposed outdoors for 36 weeks under a filter that transmitted all wavelengths in the solar spectrum developed three times as many visible checks as the control samples exposed under a filter that blocked UV and visible radiation. The checks that developed at the surface of samples exposed to the full solar spectrum were also larger than those found in samples exposed under filters that blocked UV and visible light. As pointed out by Evans et al. (2008) “this contrast in the number and dimensions of checks in samples exposed to the full solar spectrum and those in samples that were shielded from UV and visible radiation provides compelling evidence that checking of wood is increased by photodegradation of the wood surface”.  2.2.1.2 Water As mentioned above, loss of water from wood generates stresses that cause wood to check. Water also plays a major role in the weathering of wood. Many studies have concluded that water combined with other environmental factors is a significant factor leading to the deterioration of wood surfaces. For example checks expose new cell wall material to UV degradation (Rowell et al. 1981). Water can occur as gas (vapour and atmospheric humidity), liquid (rain, dew) and solid (ice, snow). Depending on its state, the impact of water on wood is different (Kühne et al. 1968). Liquid water causes leaching of low molecular weight compounds from wood, hydrolysis, swelling and shrinking, distortion, and ultimately cracking and splitting of wood (Coupe and Watson 1967, Feist and Hon 1984). According to Arnold et al. (1991), a water-spray system in a weatherometer is essential for removing UV-light-degraded wood material from exposed surfaces and obtaining surfaces similar to naturally weathered wood. Water is also capable of generating minute checks in wood, when it freezes and thaws because its volume is increased when it is solid, which causes stresses within the cell lumen and the cell wall depending on the moisture content of the wood (Rowell et al. 1981, Furuno 2001). Water can modify or accelerate chemical changes 11  and deterioration of wood initiated by UV light, such as colour, brightness and pH (Webb and Sullivan 1964, Stamm 1965, Jin et al. 1991, Turkulin and Sell 2002).  2.2.1.3 Heat Heat is not as critical a factor as UV radiation or water in causing the weathering of wood, but as temperature increases the rate of photochemical and oxidative reactions increase (Feist 1983). Hon (1981) stated that heat in combination with the presence of oxygen, moisture and impurities (sensitizers), and the topostereochemistry of cellulose played important roles in the photodegradation of wood. Heat is a catalyst for the degradation reactions of wood by light. Derbyshire et al. (1997) showed that tensile strength losses of softwood veneers exposed to artificial weathering, increased with increasing temperature. Exposure of wood to heat leads to excessive drying of the surface, which increases checking. The temperatures attained at wood surfaces exposed outdoors however, are unlikely to approach the glass transition temperature of lignin or temperatures at which significant structural degradation of wood’s chemical components occurs. Therefore it is unlikely that heat directly causes surface degradation of wood in temperate climates, but rather accelerates photooxidation and hydrolysis of wood caused by light and water, respectively (Wengert 1966).  2.2.1.4 Other Factors Other environmental factors such as wind, dust, oxygen and pollutants affect the weathering of wood. The surface of wood may be abraded mechanically by windblown particles of sand or ice crystals (Feist and Hon 1984). Small particles such as sand can become lodged in surface checks and, through swelling and shrinking, weaken fibers in contact with them (Feist 1983). Solid particles in combination with wind can also have a “sandblasting effect” (Hon 1981, Feist 1983).  12  A relatively new, anthropogenic factor influencing the weathering of wood is air pollutants particularly gases such as nitrogen dioxide, sulfur dioxide, ammonia or ozone (Hon 1983, Feist and Hon 1984). These chemicals may increase the surface degradation of wood. Williams (1987) carried out an artificial weathering test and observed higher erosion rates for western red cedar (Thuja plicata Donn. ex D. Don) samples, treated with acids of different concentrations, in comparison to untreated samples. This effect may occur outdoors when wood is exposed to acid rain (Feist 1990). It is not known if degradation caused by environment pollutants increases the susceptibility of wood to checking. In addition to the aforementioned factors and their interactive effects, oxygen influences the photodegradation of wood by participating in the formation of free radicals that depolymerize lignin and cellulose (Hon 1983). In some cases, bacteria and fungi also contribute to surface degradation and particularly the greying of weathered wood, but their effects on the structural integrity of wood is limited (Shoeman and Dickenson, 1997).  2.2.2  Wood Species The development of checks varies between wood species. Coupe and Watson  (1967) examined the location of checks in softwoods and hardwoods subjected to accelerated weathering. In softwood most checks originated in latewood, while earlywood developed larger, but fewer checks. The possible reason suggested for this was that drying caused stresses in ray and non-ray tissue that were greater in latewood than in earlywood. Another explanation could be the higher number of resin ducts in latewood because checks invariably developed in the vertical resin ducts of Douglasfir (Pseudotsuga menziesii (Mirb.) Franco), Scots pine (Pinus sylvestris L.) and radiata pine (Pinus radiata D. Don) (Coupe and Watson 1967). In support of this suggestion, Mackay (1973) concluded that surface checks are initiated in resin canals, because they are “…focal points of the tensile stresses arising from the shrinkage of the large number of surrounding latewood fibers and tracheids”.  13  Table 2.1 Location of checks in selected softwoods subjected to artificial weathering Source: Coupe and Watson 1967 (modified) Species  Location of checks Tangential surface Large checks, originating in rays, extending between adjacent tracheid walls especially in latewood.  Radial surface Longitudinal checks between adjacent tracheid walls especially in latewood.  Radiata pine (Pinus radiata)  Large checks extending from vertical resin ducts between adjacent tracheid walls. Some checks within rays.  Large checks extending from vertical resin ducts between adjacent tracheid walls. Some longitudinal checks between adjacent tracheid walls mainly in latewood.  Douglas fir (Pseudotsuga menziesii)  Long checks, originating in rays, extending between adjacent tracheid walls especially in latewood.  Checks (some long) between adjacent tracheid walls in or close to latewood.  Western red cedar (Thuja plicata)  Long checks, originating in rays, extending between adjacent tracheid walls especially in latewood  Long checks between adjacent tracheid walls almost exclusively in latewood or late earlywood. Some diagonal checking through pits between radial ray parenchyma and longitudinal tracheids.  Scots pine (Pinus sylvestris)  Source: Coupe and Watson 1967 (modified)  Coupe and Watson (1967) found that the appearance of checks in softwoods such as Scots pine, radiata pine, Douglas fir and western red cedar also varied substantially. The tangential surfaces of Scots pine following artificial accelerated weathering had “large checks, originating in rays, extending between adjacent tracheid walls especially in latewood” whereas radiata pine had “large checks extending from vertical resin ducts between adjacent tracheid walls”. “Some checks also occurred within rays”. Sell and Leukens (1971a) found that 20 species of wood showed clear differences in the way they weathered. The surface of some species became rougher and numerous checks developed, but the differences between species became less pronounced with exposure. Evans (1989b) found that macrochecks on transverse and tangential surfaces of radiata pine after 2 years of natural weathering were caused by cell wall failure and crack propagation at a microscopic level. At tangential surfaces, 14  ray tissue degraded rapidly and the resulting voids enlarged and coalesced to form macro-checks. Sandberg (1997, 1999) and Sandberg and Soderstrom (2006) exposed pine (P. sylvestris) and Norway spruce (Picea abies Karst) wood decking samples outdoors, and observed that pine developed more checks than spruce. The maximum width of checks in pine, however, was smaller than that of checks in spruce. Sandberg also suggested that pine had a greater tendency to form new cracks than spruce when boards were exposed wetting and drying cycles (Sandberg 1997, 1999). Flæte et al. (2000) examined crack formation in two different species, aspen (Populus tremula L) and spruce (Picea sp.) following exposure of wood to artificial accelerated weathering. Aspen developed a high number of relatively short checks; whereas spruce developed fewer, but longer checks. Kishino and Nakano (2004a) exposed eight tropical wood species to 600 hours of artificial weathering. They found that the development of cracks and their magnitude varied between species. They concluded that factors such as crack area on the surface played an important role in the difference between species in wettability during artificial weathering.  2.2.3  Wood Properties The properties of wood also influence the development of checks. The  influence of growth ring orientation on check formation has been extensively studied. Stamm (1965) suggested that wood used outdoor should have vertical growth rings. This “minimizes the risk of cracks as a consequence of anisotropic moisture movement”. According to Browne (1960), Stamm (1965) and Sandberg (1999), tangential surfaces of softwoods exposed to weathering have a greater number of checks per unit area than the corresponding radial surfaces. Tangential surfaces also develop wider and deeper checks than radial surfaces. The checks on tangential surfaces occur frequently in both earlywood and latewood, whereas, on radial surfaces, checks occur primarily at growth ring borders (Sandberg 1999). Sandberg  15  (1999) explained that the difference in crack susceptibility between radial and tangential surfaces was mainly the result of stresses which arise in the wood as a consequence of anisotropic gradients between the surface and sub-surface region. As mentioned earlier, tangential surfaces of softwoods check more than corresponding radial surfaces, and checks on tangential surfaces are greater in number per unit area as well as being wider than those on radial surfaces (Sandberg 1999). This phenomenon can be related to the fact that wood shrinks more in the tangential direction than radial direction, and therefore stresses and strains are greater in boards with growth rings orientated parallel to the exposed surface (Schniewind 1959). Accordingly, wood with growth rings orientated perpendicular to exposed wood surfaces, i.e. quarter-sawn boards, tends to check less than flat-sawn boards (Schniewind 1963, Sandberg 1997, 1999, Flæte et al. 2000, Yata 2001). A series of studies by Sandberg (1996, 1997, 2005) showed that boards sawn near the pith and exposed to natural weathering had shorter checks than similarly exposed boards sawn away from the pith. Flæte et al. (2000) investigated the influence of distance from the pith on checking of boards exposed to accelerated weathering and found that boards from near the bark had more severe cracks than boards from near the pith. Differences in checking of boards with similar growth ring orientation occur due to variation in wood structure. Yata (2001) studied the surface checking of naturally weathered hinoki (Chamaecyparis obtusa (Sieb. And Zucc.) Endl.), sugi, western hemlock (Tsuga heterophylla (Raf.) Sarg.) and redwood (Sequoia sempervirens (D. Don) Endl.) with different sawing patterns and growth ring orientations. In all cases, less checking developed in flat-sawn boards whose annual rings were orientated concave to the exposed surface. Sandberg and Soderstrom (2006) also found that boards exposed pith side up developed fewer and shorter checks than the boards exposed bark side up. Support for these findings was recently provided by Urban and Evans (2005) and Evans et al. (2008). They examined the surface checking of naturally weathered matched flat-sawn southern pine decking boards which were exposed with their growth rings orientated concave and convex to their  16  upper surface (Figure 2.2) 2.2).. They found that boards whose growth rings were orientated more perpendicular to the affected surface, checked less. Hence, boards with concave growth owth ring orientation developed less surface checking than boards whose growth rings were orientated convex to the upper exposed surface. Weizenegger (2006) also found that boards exposed to natural al and accelerated weathering developed fewer and smaller ch checks ecks if their rings were orientated concave rather than convex to the exposed surface.  Figure 2.2: Orientation of growth rings to the upper exposed surface of decking boards; convex orientation, left; concave orientation, right  2.3 Treatments to R Reduce Checking The surface checking of wood is mainly caused by stresses generated by shrinkage and swelling that exceeds the elastic limit of wood (Schniewind 1963, Stamm 1965, Rowell et al al. 1981, Evans 2008). Some treatments have been developed develop to prevent wood from checking including water repellents, preservatives, and mechanical treatments.. Th These will be discussed in the following sections.. 2.3.1  Water Repellents Water repellents are treatments that change the surface properties of wood so  that itt sheds liquid water. Water repellents work exceptionally well at retarding the absorption of water by the end grain, the most absorptive of the different wood 17  surfaces. Although water repellents do not stop all water absorption, they are an excellent treatment for wood used outdoors because they inhibit the absorption of liquid water, but allow the wood to dry after rain. Water repellents are usually complex blends of wax, oil, resin and solvent (Borgin and Corbett 1970a). Formulations that include a mildewcide or a wood preservative are referred to as water-repellent preservatives (Williams and Feist 1999). The oil and resin penetrates the wood surface and cures to partially seal the wood. The influence of rain and dew causing surface erosion and leaching of wood components (and the moisture level of wood in service) are reduced. Thus the risk of crack formation and biological degradation is lowered (Zahora 1991). Waxes in water repellents impart water repellency to wood and retard the ingress of moisture, but they do not prevent the wood from reaching an equilibrium moisture content with the surrounding atmosphere (Belford and Nicholson 1969). When used as a surface treatment, wax alone cannot provide lasting protection because of displacement of the wax from wood surfaces by water during weathering. Such displacement is less pronounced for waxes that contain hydrophilic groups, because the affinity of the wax for wood is better, but the hydrophobic properties of such waxes are weaker. An optimum has to be found (Borgin and Corbett 1970b). The water repellency of water-repellent preservatives also depends on the type of preservative. For example, wood treated with a water repellent ACQ preservative showed lower water repellency than wood treated with a water repellent CCA (Cui and Zahora 2000). Water-repellent additives have been used to prevent the formation of checks in CCA-treated wood during weathering (Levi et al. 1970, Zahora 1991). Zahora (1992) examined checking in Southern yellow pine boards, which had been treated with a CCA containing a water repellent (WR) emulsion additive. He measured the widths of the checks in boards after 1 year of natural weathering and found that even the third largest check for CCA-only treated boards was wider than the largest check in the CCA-WR treated boards. The water repellent additive minimized moisture gradients in treated boards during the first year of weathering and thus reduced checking.  18  A later experiment by Zahora (2000) found that the surface water repellent effect of preservative treatments containing water repellent emulsions disappeared after about three years exterior exposure. Evans et al. (2003) found that checks in CCA-wax treated radiata pine decking, following one year of natural weathering were significantly smaller in width and depth and fewer in number than those in untreated, water- and CCA-treated samples.  2.3.2  Hydrophobic Wood Preservatives There are three main types of preservatives; waterborne salts e.g., chromated  copper arsenate (CCA), preservative oils e.g., coal-tar creosote, and light organic solvent preservatives e.g., pentachrophenol. All these preservatives have been widely used by industry to treat timber (Ibach 1999). Preservatives prevent deterioration of wood by fungi and other organisms, but they are generally not formulated to protect it from physical degradation such as checking (Zahora 1991). Until recently, CCA was the most common preservative for wood used outdoors. It was used to prevent the decay of poles and piles and also for wood used for commercial and residential building purposes, e.g. for cladding, decking, etc. CCA only provided a temporary water repellent effect on southern yellow pine, which depended directly on the concentration of the treatment (Zahora 1991). As mentioned in the previous section, boards treated with CCA-wax showed smaller variations in moisture content and reduced checking. Zahora (2000) found that both CCA and CCA-WR treated boards, however, showed damage including checking caused by extensive UV exposure on their top surfaces, which promoted water-intake into boards. However, the checks in CCA-WR treated boards were scattered and usually less than 5 mm in depth, whereas checks in CCA-only treated boards penetrated much deeper into the wood (Zahora 2000). In some studies CCA has been found to increase the checking of wood, for example, radiata pine shingles exposed outdoors in Rotorua, New Zealand (Placket et al. 1984).  19  According to Bariska et al. (1988), CCA dissolved and hydrolyzed the wood substance of radiata pine in three areas: between ray and fibre tissue, between single cells in the compound middle lamella and between layers of the cell wall. This was thought to promote checking. CCA fixes or deposits material in higher concentrations in the latewood, but the damage to latewood caused by CCA is not significant since latewood is inherently strong (Bariska et al. (1988). The strength of the earlywood is not markedly affected because earlywood contains lower levels of CCA due to its reduced permeability when it is dried. However, at the interface between latewood and earlywood high concentrations of CCA in the latewood come in contact with the weaker earlywood and cause hydrolysis of the holocellulose and lignin fractions of the wood substance. Since this interfacial zone is weakened, it is the site according to Bariska et al. (1988), for the formation of delamination cracks. Evans et al. (2003) found that sawn kiln-dried CCA-treated radiata pine decking exposed to one year of weathering in South-East Australia, checked in a similar way to timber that had been pressure treated with water. They concluded that some aspect of the treatment process, possibly re-drying after pressure treatment, may have influenced checking. In comparison to untreated wood, CCA- and water-treated wood developed slightly larger, but less numerous checks. Crawford et al. (1999) treated incised and unincised softwood decking boards with CCA-Type C and weathered them for 10 years. Only unincised white pine (Pinus strobus) decks failed due to the development of transverse checks and splits around knots. Relatively new water-borne preservatives used to treat wood are Alkaline Copper Quaternary (ACQ) and copper azole (CA). ACQ and CA both prevent decay and insect attack. There are currently four standardized AWPA ACQ formulations, ACQ Types A, B, C, and D. The different formulations allow flexibility in achieving compatibility with different wood species and end-use applications. All ACQ types contain 2 active ingredients which may vary within the following limits: copper oxide (62%-71%), which is the primary fungicide and insecticide, and a quaternary ammonium compound (29%-38%), which provides additional fungicide and insect resistance properties. There are two types of copper azole: A (CBA-A), and B (CA-B).  20  Copper azole wood preservative is used for treating a variety of softwood species including southern pine, red pine (Pinus resinosa Ait), ponderosa pine (Pinus ponderosa Douglas ex Lawson), hem-fir and Douglas fir. Currently there is little information on how these newer preservatives influence the checking of wood exposed outdoors. Cui and Zahora (2000) reported that water repellents in ACQ were effective in reducing surface checking. Oil-based treatments that have been used commercially to treat wood are creosote, pentachlorophenol solutions (PCP), copper naphthenate, chlorothalonil, chlorpyrifos, oxine copper (copper-8-quinolinlate) and zinc naphthenate (Ibach 1999). Creosote and PCP are heavy oil treatments. They are water-repellent and tend to form a water impermeable oily layer on the wood surface, which reduces checking (Levi et al. 1970). Straight creosote solutions tend to be less effective at reducing checking than creosote blends (for example creosote-coal-tar solution). However, creosote preservatives tend to bleed from treated wood surfaces and penetrate wood with greater difficulty (Ibach 1999). Creosote treated messmate stringybark (Eucalyptus obliqua L’Herit) poles checked less than untreated and salt-treated poles after 5 years of weathering, irrespective of climatic conditions (Gilfedder et al. 1968). The ability of creosote to restrict checking was attributed to its tendency to inhibit rapid drying. Organic solvent-based preservatives are generally used to treat joinery and wood used above-ground (Levi et al. 1970). Copper naphthenate-based light organic solvent preservatives (LOSP), however, are considered as an alternative to CCA to overcome concerns over the use of rainwater run-off for drinking water, corrosion of fasteners, and checking. LOSPs can provide significantly improved resistance to checking (Placket et al. 1984).  21  2.3.3  Mechanical Treatments Mechanical treatments for timber were originally designed to improve  preservative penetration. These treatments have mainly been applied to poles and include incising, center boring, through-boring and kerfing. Ruddick and Ross 1979 stated that the aim of mechanical treatment of poles is to improve the penetration of chemical treatments, and also reduce drying stresses in the wood. The effectiveness of a preservative treatment depends on the thickness of the outer treated shell surrounding the untreated wood, as well as the amount of preservative introduced into the wood and the continuity of the treated shell. At the same time the treatments should reduce drying stresses and prevent checks from developing and from propagating beyond the chemically treated shell (Ruddick and Ross 1979).  2.3.3.1 Incising Incising is a mechanical pre-treatment process in which steel knives are used to make longitudinal incisions into the sawn faces of wood. The main reason for incising wood is to improve the penetration of preservatives in refractory (impermeable) species. Incising originated in Europe and was first used in the USA as a pre-treatment for Douglas fir ties and timbers in Oregon in the 1920s. Incising should be differentiated from the perforation of wood with a pattern of drilled holes or slits that run around the outside of poles with a slope of 5° (Graham 1973). Numerous tests have been conducted on the effect of incising on wood. Graham and Estep (1966) reported that incising alone did not control checking in Douglas fir spar cross-arms. The combination of a kerf and incisions, however, prevented both excessive checking and exposure of untreated wood. Graham and Estep (1966) concluded that the combination of incising and kerfing was needed to prevent small checks from exposing untreated wood in cross-arms. Henry (1970) reported that incising reduced the number of hardwood ties that developed wide checks in service. Periodic inspection was done to evaluate the effect of incising on severe checking in hardwood cross ties during 28 years service. Incised 22  ties had less severe checking than unincised ties and they also changed less in the 20 year period. Approximately 85 percent of incised ties had no checks greater than 3/8 inch wide, however, only 75 percent of unincised ties were in this condition. Incising noticeably reduced the severity of seasoning checks, but had no apparent effect on the drying rate during air seasoning (Henry 1970). Ruddick and Ross (1979) hypothesised that incising might encourage the formation of several shallow checks, rather than one major check, which could penetrate beyond the treated shell. Moreover, Graham (1973) stated that incising improved the drying process and thus reduced checking indirectly, and it only had a minor strength-reducing effect on preservative-treated poles. However, Evans et al. (1997) found that incising was ineffective at reducing checking in CCA-C treated slash pine (P. elliottii) poles after air-drying and weathering.  2.3.3.2 Through Boring and Center Boring A method of improving the penetration and retention of preservative in wooden poles is to bore a series of holes through the wood. This method is known as "through-boring" and is applied to wooden poles near the groundline (Newbill 1997). Through-boring involves drilling slightly angled holes from one face through the pole to the opposite side. Through-boring patterns typically space holes 50 to 100 mm apart across the cross-section and spaced 150 to 200 mm longitudinally (Rhatigan and Morrel 2003). This pre-treatment produces nearly complete treatment of the bored zone, largely eliminating the risk of groundline decay. Through-boring is an effective method of increasing preservative treatment and extending pole service life (Rhatigan and Morrell 2003), however, there is no information about whether it influences the checking of poles. On the other hand center boring has been reported to influence the checking of poles (Mater 1972, Graham 1973, Goodell 1991, Evans et al. 1997, 2000). Center boring is a simple technique of drilling a hole through the center of a pole or post. Mater (1972) found that center boring of untreated poles significantly reduced the number and lengths of checks that developed when poles were exposed outdoors. 23  Centre boring was also thought to have had some beneficial effect on the long term strength of poles, even though the initial strength of a centre-bored 30-foot long pole 8 inches over the ground line was about 1 % and 9.8 % weaker in bending and shear, respectively, in comparison to a solid pole. According to Graham (1973), center boring reduced the initial width of checks in CCA-treated poles, however, it had a strong negative effect on the widening of checks which more than offset its positive effect on checking. The depth of the checks in center bored poles was only reduced in untreated poles. Goodell (1991) concluded that center boring of poles led to a reduction in the width of checks formed, however, the effect was not statistically significant. Evans et al. (1997 and 2000) came to the conclusion that center boring of ACQ and CCA-C treated radiata pine and slash pine posts prevented checks from developing and from becoming wider or deeper, but it had no significant effect on the length of checks. Posts containing center bored holes with a larger diameter (Figure 2.3) developed narrower checks, but there was no significant effect of the diameter of the centre-bored hole on check depth or number (Evans et al. 1997 and 2000).  Figure 2.3: Kerfing and centre-boring treatments (Evans et al. 2000)  24  2.3.3.3 Kerfing One technique that has been shown to reduce checking of poles is kerfing (Graham and Estep 1966, Graham 1973, Helsing and Graham 1976, Ruddick and Ross 1979, Ruddick 1981, 1988, Morrell 1990, Evans et al. 1997, 2000, Kurisaki 2004). Kerfing involves making a saw cut from the periphery of the pole to its centre and from the butt to the desired point along the axis of the pole (Morrell 1990). Kerfing acts to relieve drying stresses and produces a well treated check, which shrinks and swells with moisture changes. Experiments by Graham (1973) showed that three kerfs spaced equally around posts widened on initial exposure of the posts to the weather. After a short period of time, however, only one of them continued to widen and deepen, usually to the center of the post. Thus, the author concluded that only one kerf to the center was needed to prevent checking. Graham (1973) also observed a reduction in check width and depth in CCA-treated red spruce (Picea rubens Sarg) poles that had been kerfed to their centers along their entire length. Evans et al. (2000) examined the effect of kerfing and center-boring on the checking of ACQ treated radiata and slash pine posts. They found that double kerfing was significantly more effective at preventing checks from developing and enlarging than single kerfing (Figure 2.3). For example posts containing two 30 mm and 22 mm deep kerfs had the lowest check surface area, number of checks and check volume, width, length and depth. Ruddick and Ross (1979) examined the effect of kerfing on checking of untreated, unseasoned Douglas-fir pole sections. The pole sections were kerfed full length to the pith with a 4 mm wide saw kerf. They found that the checking of the unkerfed pole sections was much more severe than that of the kerfed sections. Hence, the maximum depth and width of the largest checks was found in the unkerfed pole sections. The mean depth of the largest checks in the unkerfed poles was almost 2.4 times greater than that in the kerfed pole sections. The effect of kerfing on checking of white spruce poles was also examined by Ruddick (1981, 1988). The poles were commercially pressure treated with either ammoniacal copper arsenate (ACA) or pentachlorophenol (PCP). The average depth  25  of the largest check in the unkerfed ACA treated poles was approximately 2.2 times greater than that in kerfed poles after three and ten years of exposure. Moreover, kerfed ACA treated poles developed less checking than either unkerfed PCP or ACA treated poles. Morrell (1990) evaluated 5,000 kerfed poles of different sizes that had been in service for various periods of time. He found that if kerfing was performed prior to treatment of poles, the kerf was well treated and shrank and swelled with moisture changes and, accordingly, relieved internal stresses. In larger poles, kerfing prevented the development of deep checks extending beyond the depth of the treated shell. However, kerfing did not substantially improve the performance of smaller poles. Helsing and Graham (1976) inspected creosote or pentachlorophenol treated Douglas fir transmission poles that contained a single deep saw kerf from the butt to 1.5 m above ground or for two-thirds of the pole length after 5 to 11 years in service. They found that kerfing reduced the width of checks. However, small checks occurred at the base of those kerfs that ceased 41 mm from the center of the pole. Evans et al. (2000) found that increasing the depth of double and single kerfs reduced the number, width and depth of checks that developed when posts were exposed to natural weathering. An earlier experiment by Evans et al. (1997) found that single and double kerfing were both effective at reducing the number of checks greater than 1.0 mm wide that developed in slash pine posts during air-drying and after they were treated with CCA and exposed to 6 weeks of weathering. These treatments were also effective at reducing the number and size of checks after one year of natural weathering. Kurisaki (2004) also confirmed that single and double kerfs prevented deep checking of sugi (Cryptomeria japonica D. Don) posts. Recently the effect of kerfing on the checking of square cross-section timber has been examined. Sadoh (2001) looked at whether the checking of boxed heart timber during drying could be reduced by kerfing. Saitoh et al. (2000) also measured the number of cracks that developed in kerfed and unkerfed sugi boards during kiln drying. Both studies concluded that kerfing can reduce the checking of square crosssection timber. For example, Saitoh et al. (2000) found that checks in kerfed sugi  26  boards 114 x 114 x 3000 mm were generally small in comparison to those found in the unkerfed controls. Overall they found that 57 percent of kerfed sugi boards had no cracks, while only 6 percent of unkerfed boards were free of checks. However, Ratu et al. (2007) observed the opposite effect of kerfing on the checking of flatsawn southern pine decking board exposed to natural weathering. In their study, they applied one, two or three kerfs to depths of 10, 13, or 20 mm into the underside of 140 x 40 x 600 mm southern pine decking boards and then exposed them together with unkerfed boards to 1 year of natural weathering. They found that the kerfs applied to the decking boards became narrower particularly at the surface of the underside of the boards. This suggested that kerfing allowed the boards to flex in response to changes in their moisture content and therefore might have increased surface strain on the upper surfaces of the board, thus increasing checking. Overall they found that kerfing had no statistically significant effect (p>0.05) on the surface checking of flatsawn southern pine decking boards.  2.4 Accelerated Weathering and Checking 2.4.1  Introduction Some aspects of the weathering process occur very slowly. For example,  according to Browne (1960), wood exposed to the weather above ground erodes very slowly, approximately 6 to 7 mm in a century. Feist and Mraz (1978) found that the effect of weathering is limited to a surface layer about 2.5 mm deep and the erosion rate is slow. The average wood loss for slow-grown western red cedar material was estimated to be 16.5 mm and for fast-grown material to be 10.1 mm per 100 years. Hon and Ifju (1978) also mentioned that the erosion rate of wood exposed outdoors is slow, ~5-12 mm per century. It is time consuming and expensive to obtain information on the weathering of wood using natural exposure trials. Therefore it is important to develop methods to accelerate weathering. Artificial weathering methods are useful tools for studying the durability of wood products that will be used outdoors. They are also widely used to  27  help develop treatments to improve the resistance of wood to weathering. Because time is a crucial factor affecting costs and competitive advantage in the process of product development, it is desirable to obtain results as fast as possible. Arnold et al. (1991) mentioned that the changes that occur during natural weathering should be accelerated 5-20 times in artificial weathering devices.  2.4.2  Types of Accelerated Weathering Tests / Machines Two categories of accelerated weathering tests have been developed: (1)  outdoor; and (2) indoor. Accelerated outdoor weathering is a compromise between testing under natural and artificial conditions. The advantage of such tests is that they provide results that are close to later service conditions due to the use of natural weathering factors such as sunlight and rain water. The difference between natural and accelerated outdoor weathering is that in the latter the effects of specific environmental factors are enhanced. Factors such as period of year, climate or air pollutants, however, may confound outdoor experiments and complicate the achievement of comparable and reproducible results (Shah 1983) The American Society of Testing and Materials (ASTM) (2004a,b) provides standard methods for accelerated natural weathering tests. These methods use mirrors to concentrate sunlight or glasses to filter it (ASTM G90-98, G24-97) which increase the impact of sunlight and heat on the test material. A supplementary step that can reduce the length of testing is to expose samples at an angle to the ground in order to maximize total radiation received by samples (ASTM G7-97). Artificial weathering methods enable the user to achieve controlled and fully reproducible conditions. Despite the fact that various factors influence the weathering of wood, existing artificial weathering methods restrict the environmental factors to light, heat and water, and strongly emphasize the impact of the light source. The artificial weathering methods mainly differ in terms of their applied light source (ASTM 2004a,b) and therefore can be classified as follows: (http://www.atlasmts.com/products/laboratory-weathering-testing/)  28  -  Weathering devices employing carbon-arc lamps The light produced by the Sunshine Carbon Arc provides more UV at wavelengths below 300 nm than natural sunlight, but gives a better match than the Enclosed Carbon Arc at longer wavelengths. When used without filters for faster testing, stability rankings of some materials may be distorted when compared with outdoor testing. This technology has largely been replaced with fluorescent UV or xenon arc systems.  -  Weathering devices employing xenon-arc lamps The xenon long arc simulates UV and visible solar radiation more closely than any other artificial light source (Shah 1983). Xenon arc is a precision gas discharge lamp within a sealed quartz tube. The spectral power distribution is altered through filtering to simulate solar radiation. It is widely preferred as a light source when the material to be tested will be exposed to natural sunlight.  -  Weathering devices employing fluorescent UV lamps Tests using fluorescent lamps are useful for ranking and comparing materials exposed under specific conditions, but the comparison to service lifetime performance or correlation with outdoor exposure may not be valid. Fluorescent UV lamps that are similar in mechanical and electrical characteristics to those used for residential and commercial lighting have been developed with specific spectral distributions. These sources are incorporated into fluorescent UV condensation devices such as the UV2000. These devices may be used in tests that vary light/dark cycles, temperature, condensing humidity, water spray and irradiation. In addition to these devices some customized weathering machines have been  built to accommodate larger samples (Coupe and Watson 1967), or simulate the effect of environmental factors that are not employed in standard devices. Coupe and Watson (1967) developed an apparatus for weathering wooden samples that included water immersion, forced drying, high temperature and UV-light. Samples were arranged on a wheel and subjected to each cycle periodically. A different approach to the artificial weathering of aspen and Norway spruce was employed by Flæte et al. (2000). They  29  added freezing and conditioning to their cycle in addition to light, heat and water according to the Norwegian Standard 8140 (Norges Standardiseringsforbund 1985).  2.4.3  Correlation between Natural and Artificial Weathering Natural weathering is the result of the combined action of numerous  weathering factors as mentioned above. The effects are complicated because climate is not constant and differs from one part of the earth to another. According to Alexopolous (1992), outdoor weathering provides the most reliable data, but the differences in local and geographical climate conditions limit general applicability and comparison of test results. Accelerated weathering, however, can provide more rapid information on the resistance of a material to weathering, but it is difficult to say how the results of such tests relate to those obtained from natural weathering trials due to the complexity of natural weathering. Basically, all aging factors should be accelerated to the same extent, however, results can differ depending on whether the factors act in combination or alone (Gjelsvik 1983). Standardized equipment by its very nature cannot perfectly duplicate outdoor conditions that vary widely according to latitude, temperature, seasonal variation in the weather, sample orientation, altitude and other variables (Breard 2003). Gjelsvik (1983) pointed out … “Due to combinations of weathering effects, attempts to accelerate the individual factors to the same extent should not be over emphasized. The most important point is to develop an aging cycle and to select test samples giving a sensible correlation with practice.” The effects of accelerated testing methods can be assessed either qualitatively or quantitatively. Coupe and Watson (1967) used their device to examine the process of surface checking of naturally and artificially weathered samples. The two weathering methods were compared in terms of the locations of checks on samples. They found that the type of checking and their location in samples exposed to natural weather conditions for 6 months were identical with that observed in artificially weathered samples.  30  ASTM standards (2004a) allow correlations between natural and artificial testing periods with respect to material changes such as extent of checking or color changes. Feist and Mraz (1978) compared the erosion rates of western red cedar wood subjected to natural and accelerated weathering. They observed different erosion depths for wood subjected to natural and accelerated weathering and calculated the extent to which artificial weathering accelerated erosion. They found that the erosion rates of springwood in western red cedar samples were similar during the early stages of natural and artificial weathering. 2.4.4  Accelerated Weathering Devices and Cycles to Increase Checking of Wood Accelerated weathering has been widely used to test wood. Coupe and Watson  (1967) used two different types of weatherometers to study the weathering of wood. The first device consisted of a 22 inch diameter wheel that rotated every 2 hours. Samples on the wheel were subjected to 20 minutes of water immersion, followed by 10 minutes exposure to forced air and then exposure to temperatures of 30 – 80 ºC for 40 minutes. Finally samples were subjected to UV radiation for 10 minutes. The second device was a pole weatherometer which could test 12 poles of 5-6 inches in diameter or 4 poles 10-12 inches in diameter. The samples were exposed to a 2kW radiant heat source for 5 hours and a water spray cycle. Gjelsvik (1983) described an apparatus for the accelerated weathering of building materials and components. The device consisted of a circular chamber and three fixed boxes in which climate stresses were applied. The device ran the following cycle: radiation, spraying with demineralized water, cooling and freezing and finally thawing at room temperature. Feist and Mraz (1978) exposed western red cedar samples to a high-intensity carbon arc lamp in an enclosed chamber at 45 – 50 ºC for 5 days to evaluate the erosion of wood during weathering. They employed a cycle involving 4 h of distilled water spraying and 20 hours of light. Arnold et al. (1991) used modified fluorescent UV light and xenon arc weathering devices to compare the surface erosion of 5 different wood species. They weathered the specimens for 2400 h using a cycle 31  involving 5 h drying, 1 h water spraying in fluorescent UV weathering device or 24 h drying, 4 h water spraying in xenon arc weathering device. Plackett et al. (1992) exposed acetylated radiata pine veneer in an Atlas Ci-65 weatherometer. Their cycle consisted of 120 min of light with 18 min of distilled water spray. Kishino and Nakano (2004a) tested the changes in wettability of eight species of tropical woods during weathering using a sunshine xenon long-life weatherometer. They used the same settings as Plackett and coworkers, except the temperature was 45 ºC. This cycle was also used by Temiz et al. (2005, 2007) when they examined the weathering resistance of preservative treated Scots pine and alder (Alnus glutinosa L.) sapwood. Flæte et al. (2000) used a different approach to accelerate the surface checking of solid wooden aspen and Norway spruce boards. Their board samples were exposed sequentially for 48 hours to each of four different elements; light and heat, water spray, freezing and drying. The temperatures in the different elements of the cycle were 75 ± 5 ºC (light/heat), 18 ± 5 ºC (water spray), -20 ± 5 ºC (Freezing), and 23 ± 2 ºC (drying). The weathering cycle took 122 days to complete. This cycle was then modified by Weizenegger (2006). He subjected southern pine specimens to a five day cycle in a customized accelerated weathering device. The apparatus used halogen infra-red lamps to generate heat and used circulation fans to blow dry air across samples. Large dimension samples were placed under restraint within the apparatus and manually sprayed with water. The samples were subjected to wetting, drying and freezing cycles over a 5 day period. The maximum temperature in the apparatus during drying was set at 73 ºC and controlled using an automatic sensor. The checking of samples was assessed after each daily exposure cycle. Table 2.2 summarizes the different approaches that have been used to accelerate the weathering of wood.  32  Table 2.2 Summary of previous studies that have employed accelerated devices to examine the weathering of wood Author Coupe and Watson (1967)  Device 1. Cyclic weathering device  2. Pole weatherometer Feist and Mraz (1978)  Carbon arc chamber  Anderson et al. (1991)  Xenon arc weathering chamber  Arnold et al. (1991)  Modified fluorescent UV and Xenon arc weathering chambers Atlas Ci-65 Weatherometer  Placket et al. (1992)  Flæte et al. (2000)  Accelerated weathering apparatus  Kishino and Nakano (2004a,b)  Sunshine xenon weatherometer WEL-6XS-HCBEc-S  Temiz et al. (2005)  ASTM G53  Weizenegger (2006)  Customized weathering device  Temiz et al. (2007)  ASTM G53  Cycle  Material examined  2 hour cycle; 20 min water immersion 10 min forced air 40 min temp. 30-80 ºC 10 min UV radiation  Wooden boards  4 hour cycle; 2 kW radiant 90 ºC Water spraying 5 x 24 hours cycle 20 h light at 45-50 ºC 4 h water spraying 0, 50, 150, 300, 2400 h cycles 24 h light 4 h water spray 45 - 50 ºC temperature 2400 hour cycle 5 h drying 1 h water spraying Light  Surface checking of poles  3000 hour cycle: 2 h UV light irradiation at 340 nm, 18 min water spray 50 ºC temperature 122 day cycle: 45 min heat at 75 ºC 15 L/(m2h) wetting at 18-20 ºC freezing 20, 50, 100, 200, 300, 400, and 600 hour cycles: 2 h UV irradiation at 340 nm, 18 min water spray 45 ºC temperature 200, 400, and 600 hour cycles: 2 h UV light irradiation at 340 nm, 18 min water spray 45 ºC temperature 5 day cycle: 1 h drying 12 g water spraying 73 ºC temperature 400 and 800 hour cycles: 2 h UV irradiation at 340 nm, 18 min water spray 45 ºC temperature  Erosion of western red cedar Weathering characteristic of western red cedar, southern pine and Douglas fir Surface erosion of 4 softwood and 1 hardwood species  Acetylated radiata pine veneer  Surface checking of aspen and Norway spruce boards  Wettability of 8 tropical wood species  Surface roughness and color characteristic of treated Scots pine sapwood  Surface checking of southern pine decking boards Surface chemistry of treated Scots pine sapwood  33  2.5 Summary Based on this review of the literature, it can be concluded that surface checking of wood is mainly generated by a complex combination of weathering factors such as solar radiation, wetting and drying, and heat (Sell 1968, Chang et al. 1982, Feist 1982, 1990 and Evans 2001). During outdoor exposure, wood surfaces become rough and then check due to the development of stresses within surface and sub-surface wood layers. The stresses and strains that lead to the checking of wood occur due to shrinkage and swelling as a result of hygroscopic movement of water into the core and drying of the surface of the boards due to solar radiation. (Stamm 1965, Sell and Walchli 1969, Evans 2008). The shrinkage and swelling of wood in the tangential direction is greater than in the radial direction (Stamm 1942). Hence stresses and strains are greater at the surface of boards with growth rings orientated parallel to exposed surfaces (flatsawn boards) than in quartersawn boards. Accordingly, tangential surfaces of softwoods develop larger and greater numbers of checks than the corresponding radial surfaces (Stamm 1965, Sandberg 1999, Sandberg and Soderstrom 2006). Checks vary in size and number between soft- and hardwoods as well as within individual species. In softwoods most checks originate in latewood (Coupe and Watson 1967). According to Borgin (1971) and Evans (1989), checks develop where adjacent tissues differ in cell wall thickness or strength, e.g. at growth ring boundaries and in rays and resin canals, where stresses are concentrated, and at the interfaces between rays and tracheids. Artificial weathering methods are useful tools for studying the durability of wood products that will be used outdoors. Time is a crucial factor during the development of new products so it is desirable to obtain results as fast as possible, while simulating the degradation processes that occur during natural weathering. Methods for accelerating the weathering have been standardized by American Society of Testing and Materials (ASTM). The established weathering devices, however, do not accommodate large dimension samples such as decking boards.  34  Furthermore, they only permit vertical orientation of samples. Most of the weathering cycles that have been developed seek to minimize the surface photodegradation of materials rather than generate anisotropic distribution of stress in samples that causes the checking of wood. Consequently, a different approach to the artificial weathering of wood is needed to accelerate surface checking. Therefore this thesis focuses on developing a new type of weatherometer and optimizing weathering cycles to accelerate the checking of decking boards.  35  3  Chapter Three  CYCLE OPTIMIZATION TO ACCELERATE SURFACE CHECKING OF DECKING BOARDS IN A CUSTOMIZED WEATHERING DEVICE  3.1 Introduction As mentioned in Chapter 1, a new weathering device to accelerate surface checking of decking boards is needed. In this Chapter, the development of such a device is described. The design of the device is based on a prototype developed by Weizenegger (2006). The new device retains many of the features of the previous prototype including: (1) The horizontal arrangement of samples; (2) ability to accommodate larger samples; (3) restraint of samples; (4) fans to blow dry air across samples. The new device, however, has an automated spraying and control system, which employs a TRiLOGI program to automatically spray samples with water and control the temperature and duration of wetting and drying cycles. This chapter also describes an experiment that tested the reliability of the new device and also determined which elements of the weathering cycle were critical to the device’s function of accelerating the development of checks in decking samples. Both southern pine and western red cedar samples were exposed to accelerated weathering as part of this experiment. The specimens of both species were exposed with their growth rings orientated convex to their upper surfaces as previous research has shown that this increases the development of checks during accelerated weathering and natural exposure (Yata 2001, Urban and Evans 2005, Weizenegger 2006, Evans et al. 2008).  36  3.2 Development of the Accelerated Check Tester This section describes the materials, software and equipment used to develop a device (Accelerated Check Tester, ACT) that automates most elements of the weathering cycle used to accelerate the surface checking of decking samples. 3.2.1  Hardware Used to Construct the Accelerated Check Tester The Accelerated Check Tester was originally designed to accommodate 2  board samples that were orientated to the long axis of the device (Weizenegger 2006). This feature was retained in the improved prototype. The device also simulates factors such as heat, water and rapid drying as was the case for the previous machine. Figure 3.1 is a technical drawing of the new device. Figure 3.2 is a photograph of the device. It consists of three main parts, each of which is highlighted with a different number (1, 2 and 3).  Figure 3.1: Technical drawing of the Accelerated Check Tester  37  1  2  PLC  3  Figure 3.2: Three main parts of the Accelerated Check Tester The current version of Accelerated Check Tester consists of: (1) Upper unit containing infra-red heating elements; (2) Weathering chamber that houses the two wooden decking samples and spraying system, and (3) Lower unit containing two linear actuators and the water supply system. The upper part of the Accelerated Check Tester was originally part of an infrared curing device for wood coatings. This part of the device contains two 400 mm long quartz halogen infrared lamps arranged in parallel (USHIO America, Inc.; 220V/1850W) and also contains a digital temperature controller (Omron, E5CN) to set the time limits and range of the operating temperature for the drying (heating) cycles. The walls of the weathering chamber are made from 1 mm thick aluminum plate. The bottom of the box is made from 6 mm thick aluminum plate, and contains two openings, each measuring 155 x 472 mm to accommodate the sample holders. Small circulation fans (Xinruilian, RDM 8025S, 12V/0,11A (2x) are attached on two subtended sides of the box, both of them capable of blowing air in the same direction. These fans are connected with a ribbed rubber tube (internal diameter of 75 mm), each of which contains a polystyrene dish filled with approximately 60 g of silica-gel (W.A. Hammond Drierite Company LTD., ‘Drierite’, mash size 8) to dry air that is blown  38  across the samples. Each tube also contains a baffle that can be used to control the flow of dry air across samples. Each decking sample sits in a box, measuring 150 x 420 mm, and made from 6 mm thick clear Perspex. Immediately above each of these boxes there is a stainless steel tray that collects overspray from the water spay system and diverts it to the bottom of the box. The boxes also contain two hardwood supports measuring 25 x 25 x 120 mm. They are placed inside the holders and screwed from underneath. The decking board samples are fixed to these supports using four brown, epoxy coated, self-cutting KWIKI premium extra decking screws, 3.46 x 62.6 mm in size.  Figure 3.3: KWIKI decking screws used for fixing specimens to hardwood supports in the Accelerated Check Tester. The lower part of the weathering device supports the chamber and contains the automated spray system. This system consists of two linear actuators (Techno-Isel ELS 2) connected to size-23 stepper motors (MS23C), placed parallel to the base of the chamber. Two water filters (Pentek series) with a network of hoses are positioned underneath the chamber and linked to solenoid valves (AZCAZ 8238 series). This  39  system is connected to the actuator and then on to the sprayer. Each actuator is linked to a hook-shaped aluminum tube (Ø10 mm) which is connected to a fine spray nozzle (John Brooks Nozzle QVVA-SS). A microstepping motor driver (MD2S-P-R) stimulates the stepper motor and drives the actuator along its axis maintaining a constant speed when the nozzles spray the surface of samples with water (Figure 3.4).  Figure 3.4: Image of the ACT showing the water spraying system,(lower centre)  3.2.2  Control System In order to automatically generate wet and dry cycles in the Accelerated Check  Tester, a control system was developed. This consists of Programmable Logic Controller (PLC) linked to a PC containing a TRiLOGI program version 6.2. The code for the program was designed to run on a T100MD+PLC (Triangle Research International, Inc). This program is able to control the drying cycle (temperature) and  40  wetting cycle (volume of water applied to samples) and the timing and duration of these cycles. The control system (PLC) consists of 8 analog I/Os, 8 digital inputs and 8 digital outputs. All analog inputs are 10-bit resolution and all analog outputs are 8-bit resolution. It also has a built-in 14-pin LCD as a monitor to implement the sequence and progress of the program that controls the weathering device.  3.2.3  Ancillary Equipment The other important equipment and tools that were used during the research  were; Resistance moisture meter RDM3 (Delmhorst Instrument Co.) to measure the moisture content of samples and a balance PG5002-S (Mettler Toledo). A flatbed scanner (Microteck ScanMaker i800) was used to scan the exposed surfaces of boards prior to weathering and during the 5-day weathering cycle. As a part of this scanning process, digital images of samples were saved as Tagged Image Format Files (TIFF) within the software Adobe Photoshop (Version CS, 8.0). A UV curing machine (CC12 Conveyorized Curing System) was used to irradiate the sample boards subjected to cycle 6. The UV radiation that each sample received during one pass through the machine was measured using a UV power puck II and was as follows: UVA=1015.34 mJ/cm2, UVB= 854.28 mJ/cm2, UVC = 28.22 mJ/cm2 and Visible light = 785.57 mJ/cm2. Finally, the shape distortion of samples during weathering was measured using a purpose-built apparatus. This consisted of an aluminium table that was clamped onto two support stands, and an aluminum fence that was lapped, shimmed and mechanically clamped at 90° to the table and a digital dial gauge, Mitutoyo Model IDC1012EB with 0.01 mm resolution.  41  Figure 3.5: UV curing machine used to irradiate samples subjected to Cycle 6  3.3 Optimization of Weathering Cycle 3.3.1  Experimental Design and Statistical Analysis A factorial experiment was performed to examine the effect of different  weathering cycles and wood species (southern pine v western red cedar) on checking and distortion of board samples. Each of the weathering cycles that was tested employed different combinations of heating temperature, duration of the drying stages, freezing and exposure to UV light (see Section 3.3.3 for more detailed information on the different cycles). The experiment involved four groups (blocks) of southern pine and western red cedar boards. Within these groups the individual boards were allocated to the first experimental factor of interest namely species. Samples cut from these boards were allocated to the second experimental factor, weathering cycle. The resulting hierarchical (split-plot design) accounted for fixed effects and random variation  42  between and within boards and that occurring as a result of the sequential nature of the experiment over 24 weeks. Board samples cut from four different boards for each species provided replication at the higher level. One southern pine board and one western red cedar board were selected at random to create the first replicate or experimental block. Each board was cut into six samples which were numbered sequentially from 1 to 6. These samples were then randomly allocated to the six different weathering cycles. A pair of samples (one pine and one cedar) was randomly allocated to the two sample positions within the Accelerated Check Tester and subjected to accelerated weathering. After these samples were weathered, samples from the second board were prepared and weathered. This process was repeated for samples cut from board 3 and board 4. Analysis of variance (ANOVA) was performed to assess the fixed effects of wood species and weathering cycle and random effects on checking and shape distortion of board samples. Before the final analysis, diagnostic checks were performed to see if data conformed to the assumptions of ANOVA, i.e., normality with constant variance and independence of observation within and between samples. Statistical computation was performed using Genstat 5 (Lawes Agricultural Trust 1994, release v. 4.21). Results are presented graphically and a least significant difference bar on each graph can be used to estimate the significance of differences between individual means. In addition a table is shown at the beginning of the results section that summarises the effect of fixed factors on check and distortion parameters.  3.3.2  Preparation of Decking Boards Four southern pine decking boards measuring 2400 mm x 140 mm x 40 mm  were obtained from CSI (now Viance) in North Carolina, USA. Four western red cedar decking boards measuring 3000 mm x 140 mm x 40 mm were purchased from the Home Depot store in Richmond, British Columbia, Canada.  43  Table 3.1: Density and growth ring/cm of sample boards Southern pine  Board No  Western red cedar  Density (g/cm3)  Growth ring/cm  Density (g/cm3)  Growth ring/cm  1  0.632  2.00  0.414  7.71  2  0.482  2.86  0.412  9.43  3  0.561  2.86  0.382  8.57  4  0.608  4.86  0.375  9.71  All boards were flat sawn, but their densities and numbers of growth rings/cm varied (Table 3.1). Boards were stored in a constant climate room at 20 ± 1 °C and 65 ± 5 % relative humidity (RH) for 3 weeks. The boards were planned on all four sides using an edge planer (Martin T54) and a thickness planer (Martin T44). They were then cross-cut using an Omga RN 600 radial arm saw. Ultimately 24 decking samples for each species with a final size of 135 x 35 x 400 mm were produced. Each sample had a 3.175 mm diameter hole drilled in each corner positioned 25 mm from its side and 40 mm from its ends (Figure 3.6). The specimens were then stored in a constant climate room (20oC, 65% RH) for one week prior to weathering.  Figure 3.6: Dimensions of southern pine (left) and western red cedar (right) decking samples subjected to accelerated weathering 44  3.3.3  Accelerated Weathering  3.3.3.1 Cycle Six different weathering cycles were evaluated. Each cycle involved different combinations of weathering factors (Table 3.2). Table 3.2: Combination of weathering factors used in the different weathering cycles Cycle  Drying time  Wetting time  Temperature  Soaking  Freezing  UV  1  30 min  12 sec (12 g)  73°C  yes  yes  no  2  60 min  12 sec (12 g)  73°C  yes  yes  no  3  30 min  12 sec (12 g)  80°C  yes  yes  no  4  60 min  12 sec (12 g)  80°C  yes  yes  no  5  30 min  12 sec (12 g)  73°C  yes  no  no  6  30 min  12 sec (12 g)  73°C  yes  yes  yes  The weathering factors employed in cycle 1 were the same as the ones used by Weizenegger (2006). He used this cycle and an earlier prototype of the Accelerated Check Tester to examine the checking of decking boards with different growth ring orientations. Based on his cycle, five other cycles were developed. Cycles 2, 3 and 4 increased the severity of drying by either extending drying time (cycle 2) or increasing drying temperature (cycle 3) or increasing both drying time and temperature (cycle 4). Cycles 5 and 6 were the same as cycle 1, but they omitted the freezing step (cycle 5) or exposed samples to UV light (cycle 6). All the cycles commenced by floating decking samples for 8 hours on water at room temperature (73°C). Table 3.3 shows the steps involved in each of the six different weathering cycles.  45  Table 3.3: Steps involved in the six different accelerated weathering cycles Time  Treatment  Day 1  Cycle 1  Cycle 2  Cycle 3  Cycle 4  Cycle 5  Cycle 6  Soaking  8h  8h  8h  8h  8h  8h  Freezing  16 h  16 h  16 h  16 h  -  16 h  -  -  -  -  16 h  -  6h  6h  6h  6h  6h  6h  30 min  60 min  30 min  60 min  30 min  30 min  Temperature  73 ºC  73 ºC  80 ºC  80 ºC  73 ºC  73 ºC  Water applied  12 g  12 g  12 g  12 g  12 g  12 g  Freezing  16 h  16 h  16 h  16 h  -  16 h  Conditioning  -  -  -  -  16 h  -  UV exposure  no  no  no  no  no  yes  1.5 h  1.5 h  1.5 h  1.5 h  1.5 h  1.5 h  6h  6h  6h  6h  6h  6h  30 min  60 min  30 min  60 min  30 min  30 min  Temperature  73 ºC  73 ºC  80 ºC  80 ºC  73 ºC  73 ºC  Water applied  12 g  12 g  12 g  12 g  12 g  12 g  Freezing  16 h  16 h  16 h  16 h  -  16 h  Conditioning  -  -  -  -  16 h  -  UV exposure  no  no  no  no  no  yes  Conditioning Day 2 - 4  Weathering Drying sequence  Soaking Day 5  Weathering Drying sequence  The procedures and steps involved in weathering cycles were similar to those developed by Weizenegger (2006) and are as follows (Figure 3.7); on the first day, samples were floated for eight hours on distilled water at room temperature (20 °C). Two thirds of the boards’ substructure was submerged under water. The samples were then weighed and placed separately into sample holders. They were then fastened to the sample holders with four screws, weighed again, wrapped in plastic bags and put into a freezer at -20 °C for 16 hours. This freezing step was included in the original weathering cycle because it was hypothesized that freezing would restrict the diffusion of water from the core of the samples to their surface helping to create a steep moisture gradient between the lower and the upper sides of the specimens. The wrapping of samples restricts them from drying during the freezing step. Cycle 5 did  46  not involve freezing (as mentioned above), instead samples were simply put into a conditioning room for 16 hours. day 1 day 2,3,4; day 5 without 1.5h soaking step  ACT (6 h) cycle 1,2,3,4,5,6  Soaking step (8 h)  Wrapped samples; -Freezing step (16 h) (cycle 1,2,3,4,6) -Conditioning (16 h) (cycle 5)  UV exposure (cycle 6)  Scanning Soaking step (1.5 h)  Figure 3.7: Cyclical sequences of steps during accelerated weathering (simplified) During days 2, 3 and 4, samples and holders were weighed and placed in the Accelerated Check Tester. They were then subjected to wetting and drying cycles. The surface spraying simulates the effect of dew, which wets the surface of decking boards exposed outdoors (resulting from low night-time temperatures and condensation of the air’s moisture on boards’ surfaces). Heat generated by the ACT’s infrared lamps and desiccated air blown across the samples dries the surface of specimens.  47  After six hours of alternating wetting and drying, the specimens were taken out of the weathering device, weighed, removed from the sample holders and re-weighed. The moisture content of the unexposed, lower surface of decking board samples was measured using an electrical moisture meter (RDM3, Delmhorst Instrument Co.) to determine whether it was near fiber saturation point. Then the upper exposed surfaces of samples were scanned using a large flatbed scanner (Microteck Scanmaker i800) to generate images of checked wood surfaces. Digital images of samples showing surface checks were saved as tagged image format files (TIFF) within the software Adobe Photoshop (version CS, 8.0). Following scanning, boards subjected to cycle 6 were exposed to UV light in a UV curing machine, as described above. All boards, irrespective of cycle type, were then freely floated on fresh, distilled water for 1.5 hours to recreate a moisture gradient between the upper and lower board surfaces. At the end of the soaking cycle the moisture content of the lower side of the boards was measured, the samples weighed, wrapped in plastic bags and put into the freezer or conditioning room for 16 h. On the last day of the cycle, samples were subjected to six hours of wetting and drying followed by scanning of checked surfaces.  3.3.4  Measurement of Checking and Distortion At the end of each completed cycle, the surface checking and warping of  samples were assessed. Visible checks on the surface of samples were counted and their length and width were manually measured using a transparent Perspex ruler and an optical magnifying glass containing a calibrated graticule, respectively (Figure 3.8). The shape of checks was obtained by dividing the length of each check by their width (Christy et al. 2005).  48  Figure 3.8: Tools used for measuring the length and width of checks The distortion of the samples was measured using a purpose-built measuring table and a digital dial gauge, (Mitutoyo Model ID-C1012EB with 0.01 mm resolution), (Figure 3.9). Three parameters were used to estimate the distortion of samples: 1. Cupping (tangential bending) of samples; Cupping was measured at the centre of samples when they were held firmly against a flat aluminium measuring table. 2. Twisting (longitudinal twist); Twist was measured at the end and on the sides of the specimens in both corners. The board was placed on a measuring table with one end forced flat against the bed of the table. Twist was measured on the unrestrained end of boards. 3. Bowing (longitudinal board bending); Bow was assessed as surface and side bow.  49  Figure 3.9: Digital gauge and aluminium table used for assessing shape distortion of boards  50  3.4 Results Table 3.4 summarizes the significant effects of and interactions between experimental factors on checking and shape distortion of samples. There were significant effects of weathering cycle on checking (except check width), but there was no significant effect of cycle on distortion. Species had a significant effect on all checking and distortion parameter except twist. A complete record of the statistical analysis of data in this Chapter can be found in appendix 1. Table 3.4: Significant effects of, and interactions between, weathering cycle and wood species on checking and distortion of samples exposed in the Accelerated Check Tester to different weathering cycles Checking parameter Experimental Factors  Check number  Check length  Distortion parameter  Check length  Shape+  Cupping Twisting  Bowing  Weathering cycle (W)  *  *  NS  **  NS  NS  NS  Species (S)  ***  ***  ***  ***  ***  NS  **  WxS  NS  NS  NS  NS  NS  NS  NS  * = p < 0.05; ** = p < 0.01; *** = p < 0.001; NS = not significant (p > 0.05); + Shape is check length/check width  3.4.1  Effect of Cycle Type on Checking of Boards This experiment clearly showed that boards subjected to cycle 6 checked more  than boards exposed to the other cycles. Figures 3.10 to 3.13 show the effects of the different cycles on check parameters. Figure 3.10 shows the total number of checks formed at the surface of board specimens that were subjected to the different cycles. The number of checks that formed at the surface of samples subjected to cycle 6 was significantly (p<0.05) greater than the number of checks that formed in samples subjected to the other cycles. On average, 43.7 checks developed at the surface of samples exposed to cycle 6. The comparable values for samples exposed to cycles 1, 2, 3, 4 and 5 were 24.3, 21.3, 18.8, 26.4, and 28.0, respectively. There was no  51  significance difference (p>0.05) in the number of checks that formed in samples subjected to cycles 1, 2, 3, 4, and 5. P = 0.020  Total check number (n)  50 43.7  45  l.s.d. = 13.7  40 35 30 25  26.4  28  24.3 21.3 18.8  20 15 1  2  3  4  5  6  Cycle type  Figure 3.10: The effect of cycle type on the total number of checks (averaged across southern pine and western red cedar samples) Figure 3.11 shows that checks in samples subjected to cycle 6 were significantly (p<0.05) wider than those in samples subjected to the other cycles. There was no significant difference (p>0.05) in the width of checks in samples subjected to cycles 1, 2, 3, 4, and 5. p = 0.067  Total check width (mm)  3  2.5  2  2.54  l.s.d. = 0.66  1.91  1.81  1.8 1.66 1.51  1.5  1 1  2  3  4  5  6  Cycle type  Figure 3.11: The effect of cycle type on total check width (averaged across southern pine and western red cedar samples) 52  Figure 3.12 shows the effect of cycle type on the total length of checks in samples subjected to the different cycles. The length of checks at the surface of samples subjected to cycle 6 was significantly (p<0.05) higher than the length of checks in samples subjected to the other cycles. The total length of checks in samples subjected to cycle 6 was almost double that of samples subjected to the other cycles. p = 0.005  Total check length (mm)  1000  899  900  l.s.d. = 250  800 700 600  563  500  439  452  429 381  400 300 1  2  3  4  5  6  Cycle type  Figure 3.12: The effect of cycle type on total check length (averaged across southern pine and western red cedar samples)  p < 0.01  Total check length/width ratio  20000  16048  l.s.d. = 5579  15000  10000  9097 7998 6403  5835 4784  5000  0 1  2  3  4  5  6  Cycle type  Figure 3.13: The effect of cycle type on check length/width ratio (averaged across southern pine and western red cedar samples) 53  Figure 3.13 shows the effect of cycle type on the checks length/width ratio (shape) in samples subjected to the different cycles. This parameter was significantly higher (p<0.01) in specimens subjected to cycle 6 than in samples exposed to the other cycles.  Cycle 1  Cycle 2  Cycle 3  Cycle 4  Cycle 5  Cycle 6  Figure 3.14: Scanned images of southern pine samples after exposure to different weathering cycles  Figure 3.15: Scanned images of western red cedar samples after exposure to different weathering cycles  54  Scanned images of samples subjected to the different cycles confirm that checking was more severe in samples subjected to cycle 6 than in samples subjected to the other cycles (Figures 3.14 & 3.15). This effect was observed in both southern pine (Figure 3.14) and western red cedar (Figure 3.15) samples although it is difficult to see the checks in the western red cedar samples. 3.4.2  Effect of Wood Species on Checking of Boards Checking was much more pronounced in southern pine samples than in  western red cedar samples. Figure 3.16 shows the total numbers of checks in the two wood species (averaged across cycle type). The total number of checks was significantly (p<0.001) greater in southern pine samples (42.2) than in western red cedar samples (11.9). Generally, checks in southern pine samples first developed at the centre of boards. These checks became visible on day 2 of the cycle and then became wider and longer with time during the exposure cycle. The check in the centre of boards was usually the largest check at the end of the cycle. Checks also developed in other parts of southern pine samples after day 2 of the cycle. Interestingly, there was no particular pattern of checking in western red cedar samples. The first visible check developed in different areas of the boards.  p < 0.001  50 Total check number (n)  42.2 40  l.s.d. = 9.1 30 20 11.9 10 0 Southern pine  Western red cedar Wood species  Figure 3.16: The effect of wood species on check number (averaged across cycle type) 55  The checks that developed in southern pine samples subjected to the different cycles were significantly wider (p<0.001) than those in similarly exposed western red cedar samples (Figure 3.17). p < 0.001  Total check width (mm)  4 2.93  3  l.s.d. = 0.51  2  1  0.82  0 Southern pine  Western red cedar Wood species  Figure 3.17: The effect of wood species on total check width (averaged across cycle type)  Figure 3.18 shows the total length of checks that formed in southern pine and western red cedar samples (averaged across cycle type). p < 0.001  1000 Total check length (mm)  872  800  l.s.d. = 153.7  600  400 182  200  0 Southern pine  Western red cedar Wood species  Figure 3.18: The effect of wood species on total check length  56  The total length of checks that developed in southern pine specimens was 872 mm, which was highly significantly greater than that in western red cedar samples (182 mm). The check length/width ratio parameter (shape) was also significantly higher (p<0.001) in southern pine samples than in western red cedar samples (Figure 3.19). p < 0.001 16000 Total check length/width ratio  13828  12000  l.s.d. = 2937  8000  4000  2894  0 Southern pine  Western red cedar Wood species  Figure 3.19: The effect of wood species on check length/width ratio (shape)  3.4.3  Effect of Cycle Type on Shape Distortion of Boards Analysis of variance revealed that there was no significant effect (p>0.05) of  cycle type on the shape distortion of boards. Nevertheless, the effects of the different cycles on cupping, bowing and twisting are shown in Figure 3.20, 3.21 and 3.22, respectively. The cupping of samples subjected to cycles 1, 4, 5 and 6 were close to 1 mm while that of samples subjected to cycles 2 and 3 shows was smaller (Figure 3.20).  57  p > 0.05  Cupping (mm)  2  l.s.d. = 0.504  1.5  1.034  1.094  1.095  1.099  4  5  6  1 0.81 0.685  0.5 1  2  3  Cycle type  Figure 3.20: The effect of cycle type on cupping (averaged across southern pine and western red cedar samples)  The bowing of samples subjected to cycles 1, 2 and 3 was similar, however samples subjected to cycle 4, 5 and 6 bowed more. There was large variation in the bowing of replicate board samples and this is reflected in the large error (LSD) bar in Figure 3.21. Figure 3.22 shows the twist of board samples subjected to the different cycles. Twist was greatest in samples subjected to cycle 3 (0.46 mm), whereas there was little twist in samples subjected to cycle 1. p > 0.05 1.6  Bowing (mm)  1.4 1.22  l.s.d. = 0.812  1.2  1.13 0.95  1.0 0.8 0.6  0.61  0.61  1  2  0.66  0.4 3  4  5  6  Cycle type  Figure 3.21: The effect of cycle type on bow (averaged across southern pine and western red cedar samples) 58  p > 0.05 0.6  Twisting (mm)  l.s.d. = 1.07  0.46  0.4 0.3  0.3 0.2  0.19  0.2 0.02 0 1  2  3  4  5  6  Cycle type  Figure 3.22: The effect of cycle type on twist (averaged across southern pine and western red cedar samples)  3.4.4  Effect of Wood Species on Shape Distortion of Boards Analysis of variance revealed that there was a significant (p<0.001) effect of  species on cupping and bowing, but not twist.  Figures 3.23 to 3.25 show the  differences in shape distortion of southern pine and western red cedar samples subjected to the different cycles. Figure 3.23 shows that southern pine samples cupped more than western red cedar samples. The cupping of pine boards was twice that of western red cedar boards. Images in Figure 3.26 illustrate the differences in shape distortion of southern pine and western red cedar samples.  59  p < 0.001  1.5 1.301  Cupping (mm)  1.2  l.s.d. = 0.21  0.9 0.638  0.6  0.3  0.0 Southern pine  Western red cedar  Wood species  Figure 3.23: The effect of wood species on cupping (averaged across cycle type) Figure 3.24 shows the difference in bowing of southern pine and western red cedar samples. The bowing of southern pine samples (1.24 mm) was significantly greater (p<0.01) than that of western red cedar samples (0.49 mm).  p < 0.01  1.5 1.24  Bowing (mm)  1.2  l.s.d. = 0.51  0.9 0.6  0.49  0.3 0.0 Southern pine  Western red cedar Wood species  Figure 3.24: The effect of wood species on bow (averaged across cycle type) There was no significant difference (p > 0.05) in the twist of southern pine and western red cedar samples subjected to the 6 different cycles. Pine boards twisted  60  more than cedar boards, but the difference was small (Figure 3.25). Figure 3.26 shows the shape distortion of a southern pine and western red cedar board after accelerated weathering.  p > 0.05  0.25  Twisting (mm)  0.2  l.s.d. = 0.53 0.17  0.15 0.11 0.1 0.05 0 Southern pine  Western red cedar Wood species  Figure 3.25: The effect of wood species on twist (averaged across cycle type)  Figure 3.26: Shape distortion of western red cedar (left) and southern pine (right) samples after weathering (Cycle 4)  61  3.5 Discussion The ability of the Accelerated Check Tester to accelerate checking was fully examined during the experiment described in this chapter that examined the effect of cycle type and species on checking of board samples. In order to complete this experiment the Accelerated Check Tester ran continuously for 6 hours each day for the different cycles. A total of 24 weeks was required to complete the experiment. During the experiment the device consistently maintained all the weathering factors programmed for the 6 different cycles and ran until the end of the experiment without breaking down. In addition, all of the supporting systems such as the heating, spraying and control systems worked well. Experimentation in this chapter confirmed that the Accelerated Check Tester was able to cause the surface checking of decking board samples. All of the specimens placed in the chamber checked. Interestingly, the checks mostly developed during the early stage of the exposure cycles. The checks then gradually formed over the entire surface of the boards with increasing exposure. The timing of water spraying and different temperatures used in the cycles did not significantly affect the development of checks. For example, in Cycle 1 (30 minutes drying at 73 ºC) the total number of checks that developed in samples (24.3) was higher than that of samples subjected to a cycle that dried boards for 60 minutes at 73 ºC (Cycle 2, 21.3). Furthermore, when the temperature was increased to 80 ºC there was no significant increase in check numbers or their sizes. The freezing treatment that was part of the original cycle developed by Weizenegger (2006) also had no significant effect on the development of checks. For example, in Cycle 5 no freezing treatment was applied and instead the samples were put into a conditioning room (20ºC, ± 65% r.h.). The samples subjected to this cycle had almost the same number of checks as samples subjected to Cycles 1 to 4, which all employed freezing. Thus, there is little indication that freezing affected the checking of boards. In contrast, samples subjected to the cycle that included UV exposure (Cycle 6) developed significantly more and larger checks than samples subjected to the other  62  cycles. UV radiation affects wood’s structural constituents to different degrees depending on the ability of woods’ chemical constituent to absorb UV light. These chemical reactions are accelerated by heat (Futó 1974). Sell and Leukens 1971a stated that erosion of wood including checking is closely linked with the intensity of the UV absorption and the photochemical resistance of the wood constituents. Findings here confirm those of Evans et al. (2008) that exposure to UV light accelerates the checking of wood. The results in this chapter clearly show that southern pine and western red cedar samples check differently when exposed to accelerated weathering. Pine developed higher numbers of checks and they were also wider and longer. Western red cedar developed fewer and smaller checks than pine. The mean numbers of checks in southern pine were 3 times greater than those in western red cedar. Moreover, the length of the longest check in pine was four times greater than that of the longest check in western red cedar. The difference in the checking of the two species is probably related to the lower overall shrinkage and higher dimensional stability of western red cedar compared to southern pine. Distortion in the form of cup, bow and twist developed when samples were exposed in the Accelerated Check Tester. When wood is dried, forces arise as a consequence of shrinkage in the wood material (Archer 1987). These forces can either interact with or conteract each other and determine the shape of the wood after drying. Results in this chapter showed that the different cycles had no significant effect on the shape distortion of boards. Southern pine samples, however, were less resistant to shape distortion than western red cedar samples when exposed to accelerated weathering. For example, cupping and bowing of pine boards was twice that of the cedar boards. Again, the lower overall shrinkage and higher dimensional stability of western red cedar compared to southern pine might explain these differences. The finding that checks developed at the beginning of the exposure cycle was unexpected. One possible reason for this may lie in the high moisture gradients that developed in specimens at the start of the weathering cycles. Prior to exposure the  63  samples were soaked in distilled water and when they were placed in the device the temperature rapidly rose to 73 ºC. It should also be noted that during the soaking process on the first day, the upper surface of the specimens stays dry and only the lower side of the board gets soaked with water. This causes a high moisture gradient to develop between the upper surface and the rest of the board, which would lead to the early development of stresses that cause checking.  3.6 Conclusions The purpose of the experiment described in this chapter was to test the reliability of the Accelerated Check Tester (ACT) and determine the elements of weathering cycles that significantly affect the checking of decking samples exposed in the device. From the results in this chapter it can be concluded that the device was able to cause the surface checking of decking samples. During the experiment, this new prototype device ran trouble free. UV exposure had a very large effect on checking, but freezing and increases in the severity of drying did not significantly increase checking. Hence, a weathering cycle to accelerate checking should include UV exposure, as well as wetting and drying cycles, but it is not necessary to include a freezing step.  64  4  Chapter Four  COMPARISON OF CHECKING IN SOUTHERN PINE SAMPLES SUBJECTED TO ACCELERATED WEATHERING AND NATURAL EXPOSURE  4.1 Introduction Chapter 3 described the development of a weathering device designed to accelerate the surface checking of decking boards and an experiment to determine which elements of a weathering cycle were critical to accelerating checking. The ultimate aim of the research in this thesis is to develop and to test a weathering device that can accelerate checking, but at the same time obtain results that are representative of those obtained from natural weathering trials. Therefore, an additional experiment was carried out to compare the checking of artificially weathered boards with naturally weathered boards. This experiment also obtained information on the extent to which the device accelerates surface checking of boards compared to boards exposed to natural weathering. There is no precise information in the literature on the optimum degree of acceleration for materials exposed in weathering devices. Arnold et al. (1991) mentioned that material changes under artificial weathering should be 5 - 20 times greater than those obtained during natural weathering. According to Gjelsvik (1983), the acceleration factor should be from 12 to 15 times the average for natural exposure. As a general principal, however, the greater the degree of acceleration during artificial weathering the less the material changes will resemble those obtained in-service. Hence, an important aim of the research in this chapter was to determine the authenticity of results obtained by testing samples in the weathering device and the degree of acceleration obtained.  65  The materials and equipment used were generally the same as those described in Chapter 3. The test cycle used for artificial weathering was similar to Cycle 6 described in Chapter 3, with the following differences: 1. the amount of water sprayed every 30 minutes was increased from 12 to 18 grams; 2. a 1.5 hours wetting cycle was applied at the end of the regular cycle when the samples were still under restraint. This wetting cycle also employed 18 grams water sprayed every 10 minutes, but a lower temperature (15°C) was used to prevent the surface of boards from drying out and to maintain the boards in a fully saturated condition. This part of the cycle simulates the condition of the boards when they are exposed outdoors on a rainy day and replaces the 1.5 hours of soaking that was used previously in Cycle 6. Decking boards exposed outdoors were cut from the same parent material as that used to obtain samples exposed to artificial weathering. Furthermore, the samples exposed to artificial and natural weathering were the same size. One of the samples from each pair, which was allocated to artificial or natural weathering, was longitudinally kerfed on its underside. The kerfing treatment was used to examine its effectiveness at preventing the checking and cupping of boards during weathering. As in the experiment described in Chapter 3, the checks formed on the surface of boards during weathering were manually counted and the length and width of the checks were measured. In addition, the distortion of boards following weathering was also assessed.  4.2 Experimental Design and Statistical Analysis A factorial experiment was performed to examine the effect of the different weathering treatments (natural and accelerated) and mechanical treatment (kerfed and unkerfed) on the checking and distortion of southern pine board samples. Samples cut from five different southern pine boards provided replication at the higher level. One southern pine board was selected at random to create the first replicate or experimental block. Each board was cut into 2 samples and these were allocated to natural or accelerated weathering. Each sample was then cut in two to  66  obtain 4 specimens. One specimen from each of the two samples was kerfed and the other acted as the unkerfed control. Each pair of specimens (one kerfed and one unkerfed) from the five different boards allocated to the natural weathering treatment was fixed to separate weathering racks and exposed outdoors. A matching pair of specimens (one kerfed and one unkerfed) from each block was subjected to accelerated weathering. These specimens were randomly allocated to the two sample positions within the weathering device. After specimens from board 1 (block 1) were weathered, samples from the second board were prepared and weathered. This process was repeated for specimens from boards 3, 4, and 5. Analysis of variance (ANOVA) was performed to assess the effects of the two experimental factors (weathering type and kerfing) and random effects on the checking and distortion of specimens. Data for average check length and width were unbalanced because boards contained unequal numbers of observations (checks). For such data, analysis was by weighted least squares following estimation of variance components by Restricted Maximum Likelihood (REML). This methodology was chosen because it is particularly suitable and efficient for modeling of unbalanced data with both fixed and random effects (Searle et al. 1992). The change in deviance statistic was used to measure the importance of fixed effects (weathering type and kerfing), which is referred to an appropriate x2 distribution to obtain significance levels (p-values). Statistical computation was performed using Genstat 5 (Lawes Agricultural Trust 1994, release v. 4.21). Before the final analysis, diagnostic checks were performed to see if data conformed to the assumptions of ANOVA, i.e., normality with constant variance. Results are presented graphically and 95% confidence intervals or a least significant difference bar on each graph can be used to estimate the significance of differences between individual means. A complete record of the statistical analysis of data in this chapter can be found in Appendix 2.  67  Figure 4.1: Allocation of specimens to experimental factors  4.3 Materials and Methods 4.3.1  Sample Preparation Five southern pine decking boards measuring 2400 mm x 140 mm x 40 mm  were obtained from CSI (now Viance) in North Carolina, USA. The rate of growth and wood density of the five different boards varied and is shown below (Table 4.1). All boards were stored in a constant climate room at 20 ± 1°C and 65 ± 5 % relative humidity (RH) for 3 weeks. The boards were planed on all four sides using an edge planer (Martin T54) and a thickness planer (Martin T44). They were then cross-cut into two samples (2 x 1200 mm) using an Omga RN 600 radial arm saw. These two decking board samples were then cross-cut to produce two 40 mm long decking board specimens.  68  Table 4.1: Wood characteristics of southern pine samples Wood Characteristics Board  Growth rings/cm  % latewood  Basic Density (g/cm3)  1  2.5  35.71  0.697  2  2.75  28.57  0.542  3  2.5  32.85  0.558  4  3.0  42.86  0.587  5  3.0  31.42  0.634  One specimen cut from each of the two matching board samples was kerfed longitudinally using an Altendorf F45 table saw. The depth of the kerf was 17.5 mm (50% of the thickness of the board) and the width of the kerf was 3 mm. Overall 20 decking samples with a final size of 135 x 35 x 400 mm were produced (10 kerfed and 10 unkerfed). Each specimen had a 3.175 mm diameter hole drilled in each corner positioned 25 mm from its side sides and 40 mm from its ends. The specimens were then stored in a constant climate room (20oC, 65% RH,, as above) for one week prior to weathering. Figure 4.2 shows the board samples prior to weathering.  Figure 4.2: Kerfed and unkerfed samples prior to weathering  69  4.3.2  Decking Racks Five custom-made horizontal decking racks were built out of untreated yellow  cedar (Chamaecyparis nootkatensis (D. Don) Spach) (frame) and maple (Acer sp.) (legs) serving as supports for the decking board specimens (Figure 4.3).  Kerfed  Unkerfed  Figure 4.3: Kerfed and unkerfed decking board specimen (arrowed) fastened to a decking rack Each rack contained two decking board specimens (kerfed and unkerfed). The specimens were exposed with their growth rings orientated convex to their upper surfaces (Figure 4.3). Each specimen was fastened to the substructure of the rack using four, green, epoxy-coated decking screws. Two yellow cedar boards were fastened to each rack, at the margins of the southern pine specimens, to ensure that the decking specimens were not directly exposed to solar radiation and excessive drying at their edges.  4.3.3  Weathering Schedules  4.3.3.1 Natural Weathering The five decking racks, each containing two southern pine specimens, were exposed outdoors in the compound of the Centre for Advanced Wood Processing at University of British Columbia in Vancouver (49°.11’ N, 123°.10’ W). Prior to  70  weathering, the upper surface of the specimens were scanned to produce digital images of the surface of specimens.  Figure 4.4: Decking board specimens exposed to natural weathering  On the first day of the exposure trial the specimens were fastened to the decking racks. They were then exposed outdoors for 5 days. The specimens were then removed from the racks, weighed and their moisture contents measured using an electrical moisture meter (Delmhorst RDM3). The specimens were placed in a conditioning room at 20 ± 1 °C and ± 5% RH, for 2 days, re-weighed and their moisture contents re-measured. The exposed surfaces of the specimens were scanned using a flatbed scanner and digital images of the boards were saved as TIFF files. The distortion of the specimens was also measured (see below, 4.3.3.2). These steps were repeated every week until 26 weeks had elapsed.  71  4.3.3.2 Accelerated Weathering Half of the board specimens were subjected to artificial weathering. The weathering cycle that was used was based on cycle 6 described in Chapter 3, but with some modifications. This new cycle consisted of two parts: (1), 6 hour dry cycle, which was the same as that used in cycle 6, except the amount of water sprayed on to samples every 30 minutes was increased from 12 to 18 g; (2), an additional wet cycle during which specimens were sprayed with 18 g of water every 10 minutes for 1.5 hours at ambient temperature (15°C). Instead of floating specimens on distilled water as described in Chapter 3, board specimens were immediately placed in the weathering device on day 1. They were then subjected to a dry cycle for 6 hours, followed by UV exposure in the UV curing machine, and then subjected to the wet cycle for 1.5 hours. It was clear from results in Chapter 3 that freezing had no significant effect on accelerating surface checking of decking boards. Therefore this step was omitted from the cycle used for the experiment described here. The specimens remained in the weathering chamber at the end of the wet cycle for 16 hours (overnight) until the dry cycle began on the next day.  Table 4.2: Steps involved in the artificial weathering cycle Time  Cycle  Conditions Drying sequence  Dry Cycle (6 h)  Temperature  73°C  Application of water  18 g  UV exposure Day 1 - 4  12 sec  Scanned Wet Cycle (1.5 h)  Day 5  30 min  Dry Cycle (6 h)  Wetting sequence  10 min  Temperature  15°C  Application of water  18 g  Drying period  30 min  Temperature  73°C  Application of water  18 g  Scanned  72  Prior to weathering on the first day, the specimens were scanned to obtain original images of unweathered surfaces. The specimens were then weighed and placed separately into sample holders, fastened to the sample holders using four greenepoxy coated decking screws, weighed again and placed in the Accelerated Check Tester. They were then subjected to the dry cycle of the accelerated weathering regime, removed from the device, weighed and exposed to UV radiation. The samples were then re-weighed and the moisture content of the unexposed, lower surface of specimens was measured using an electrical moisture meter (Delmhorst RDM3). The upper exposed surfaces of specimens were scanned to generate images of checks that had developed. The boards were then fastened again into sample holders and placed in the checking device. They were then subjected to the wet cycle, as described above. These steps were repeated for days 2, 3, 4 and 5, except at the end of day 5, no wet cycle was applied. The surface checking and shape distortion of boards were manually assessed at the end of each day’s accelerated weathering cycle, and after 26 weeks of natural exposure. The visible checks on the surface of specimens were counted and their length and width measured as described in Chapter 3. The cupping, bowing and twisting of specimens after weathering were measured using a purpose-built measuring table and a digital dial gauge (Mitutoyo Model ID-C1012EB with 0.01 mm resolution), also as described in Chapter 3.  73  4.4 Results 4.4.1  Checking Analysis of variance (ANOVA) revealed that there were no significant  differences (p>0.05) in the number, length and width of checks that formed in specimens subjected to 26 weeks (130 days) of natural weathering or 5 days of artificial weathering. Figure 4.5 shows the total number of checks that formed at the surface of the specimens subjected to natural or artificial weathering. The number of checks in specimens exposed to 5 days of artificial weathering (kerfed = 97.6; unkerfed = 105) was nearly the same as that formed in specimens subjected to 26 weeks of natural weathering (kerfed = 100.4; unkerfed = 109.8). The number of checks in kerfed boards subjected to artificial or natural weathering was slightly lower than that in similarly exposed unkerfed boards, but the difference was not statistically significant (p>0.05). p= 0.912  150  Check number per specimen  140 130  l.s.d = 52.7  120 105  110 100  97.6  109.8 100.4  90 80 70 60 50  Kerfed  Unkerfed  Artificial  Kerfed  Unkerfed  Natural  Figure 4.5: The total number of checks in specimens subjected to artificial or natural weathering Figure 4.6 shows the total length of checks in specimens subjected to artificial or natural weathering. The length of checks in specimens exposed to artificial weathering was slightly lower (kerfed = 97.6mm; unkerfed = 105 mm) than that in  74  samples subjected to natural weathering (kerfed = 100.4 mm; unkerfed = 109.8 mm). However, the difference was statistically insignificant (p>0.05). Similarly, there was no significant (p>0.05) effect of kerfing on the total length of checks in specimens.  p = 0.748  2400 2200  Total check length  2043  l.s.d.=1045  2072 1958  2000 1819 1800 1600 1400 1200 1000  Kerfed  Unkerfed  Artificial  Kerfed  Unkerfed  Natural  Figure 4.6: The total length of checks in specimens subjected to artificial or natural weathering Figure 4.7 shows the differences in total width of checks in kerfed and unkerfed specimens subjected to artificial or natural weathering. Again there is no significance effect (p>0.05) of weathering type or kerfing on total check width, even though the total check width in specimens subjected to artificial weathering was slightly higher (kerfed = 7.41 mm; unkerfed = 7.94 mm) than that in specimens subjected to artificial weathering (kerfed = 6.47 mm; unkerfed = 7.61 mm).  75  p = 0.599  10.0  Total check width  l.s.d. = 3.39  8.0  7.94 7.61  7.41 6.47  6.0  4.0  Kerfed  Unkerfed  Artificial  Kerfed  Unkerfed  Natural  Figure 4.7: The total width of checks in specimens subjected to artificial or natural weathering REML variance component analysis revealed that there were no significance differences in the average length and width of checks that formed in samples subjected to artificial or natural weathering. Figure 4.8 contains box plots showing the average length of checks in kerfed and unkerfed specimens subjected to artificial or natural weathering. The average check length in kerfed and unkerfed boards was similar, irrespective of type of weathering exposure, although unkerfed boards contained some checks that were much longer than those found in kerfed boards. The average length of checks in kerfed boards exposed to artificial weathering was 19.08 mm whereas it was 19.16 mm for specimens exposed to natural weathering. The average check length in unkerfed boards was 20.04 and 18.99 mm for specimens exposed to artificial and natural weathering, respectively (Figure 4.8).  76  p = 0.616  Average check length (mm)  200  175  150 40 35 30 25 20 15 10 5 0  Kerfed  Unkerfed Artificial  Kerfed  Unkerfed Natural  Figure 4.8: The average check length of specimens subjected to artificial or natural weathering  The distribution of check width in kerfed and unkerfed boards subjected to weathering is shown in Figure 4.9. The checks that developed in boards were mainly 0.05 and 0.1 mm wide. For example, the total number of checks that developed on all kerfed boards subjected to artificial weathering was 488, and 394 of them were 0.05 mm wide and 57 of them were 0.1 mm wide. Figure 4.9 also shows that artificial weathering created some checks over 5 mm wide which is wider than any of the checks that developed during natural weathering. Accordingly, analysis of variance revealed that the average width of checks in boards subjected to artificial weathering was slightly higher than that in boards exposed to natural weathering, but the difference was not statistically significant (p>0.05). The average width of checks in kerfed and unkerfed boards exposed to artificial weathering was 0.078 and 0.077 mm, respectively. Comparable figures for boards exposed to natural weathering were 0.064  77  mm (kerfed) and 0.07 mm (unkerfed), respectively. Figure 4.10 shows the scanned images of boards after 5 days of artificial weathering and 26 weeks of natural weathering. p = 0.382  1.0  2  check width (mm)  0.9 0.8  1  0.7  1  0.6  2  1  1  0.5  3  5 2  0.4  6  3  0.3  10  8  5  10  0.2  17  27  23  23  0.1 0.0  57  42  51  43  394  438  425  458  Kerfed  Unkerfed  Kerfed  Unkerfed  Artificial  Natural  Figure 4.9: Distribution of check widths in specimens subjected to artificial or natural weathering 1 WEEK  ARTIFICIAL WEATHERING  KERFED  NO-KERF  26 WEEKS  NATURAL WEATHERING  KERFED  NO-KERF  Figure 4.10: Images of kerfed and unkerfed boards (block 5) after 5 days of artificial weathering or 26 weeks of natural weathering 78  Figures 4.11 and 4.12 show the number of checks and total length of checks that developed in samples during natural weathering. The dashed line on each graph represents the number or total length of checks in specimens exposed to 5 days of  2.2  158.5  2.0  100.0  1.8  63.1  1.6  39.8  1.4  25.1  1.2  15.9  1.0  10.0  0.8  6.3  0.6  4.0  Check numbers (n)  Check numbers (Log10 n)  artificial weathering.  2.5  0.4 0  2  4  6  8  10  12  14  16  18  20  22  24  26  28  Exposure to natural weathering (weeks)  3.4  2511.9  3.2  1584.9  3.0  1000.0  2.8  630.9  2.6  398.1  2.4  251.2  2.2  158.5  2.0  100.0  Total check length (mm)  Total check length (Log10 mm)  Figure 4.11: Check number in specimens exposed to natural weathering compared to the number (dashed line) that developed in specimens exposed to 5 days of artificial weathering  63.1  1.8 0  2  4  6  8  10  12  14  16  18  20  22  24  26  28  Exposure to natural weathering (weeks)  Figure 4.12: Total length of checks in specimens exposed to natural weathering compared to the total length of checks (dashed line) that developed in specimens exposed to 5 days of artificial weathering 79  The data in Figures 4.11 and 4.12 were obtained by counting and measuring the checks that could be seen on digital images of weathered specimens. The results show that the number of checks and the total length of checks in specimens exposed to 5 days artificial weathering was very similar to those that developed during 16 weeks of natural exposure. Table 4.3: Average check length of deck board samples subjected to artificial weathering Exposure (days)  F-value  Average check length (log10, mm)* Kerfed  Unkerfed  LSD†  1  0.308  1.312 (20.5)  1.356 (22.7)  0.073  2  0.797  1.318 (20.8)  1.332 (21.5)  0.099  3  0.981  1.310 (20.4)  1.311 (20.5)  0.078  4  0.528  1.324 (21.1)  1.288 (19.4)  0.103  5  0.376  1.287 (19.4)  1.244 (17.5)  0.085  *Check length in mm in parentheses; † LSD = Least significant difference (log10)  Table 4.4: Average check length of deck board samples subjected to natural weathering Exposure (weeks) F-value Average check length (log10, mm)* Kerfed  Unkerfed  LSD†  2  0.366  1.443 (27.7)  1.396 (24.9)  0.093  4  0.486  1.359 (22.8)  1.374 (23.6)  0.043  8  0.538  1.370 (23.4)  1.358 (22.8)  0.029  12  0.727  1.294 (19.7)  1.288 (19.4)  0.032  16  0.864  1.312 (20.5)  1.320 (20.9)  0.088  20  0.578  1.285 (19.3)  1.315 (20.7)  0.100  26  0.544  1.259 (18.2)  1.282 (19.1)  0.069  *Average check length in mm in parentheses; † LSD = Least significant difference (log10)  80  Tables 4.3 and 4.4 show the average length of checks in board samples subjected to artificial and natural weathering. Results in Table 4.3 shows that the average check length of kerfed and unkerfed boards subjected to artificial weathering was nearly the same and the difference between the two was statistically insignificant (p>0.05). In the first 3 days of artificial weathering the average length of checks in unkerfed boards was slightly higher than that of kerfed boards, but the checks became shorter on average with increasing exposure. The same effect also occurred in boards subjected to natural weathering (Table 4.4). Table 4.4 also shows that there were no significant differences in the average check length of kerfed and unkerfed board subjected to natural weathering.  4.4.2  Distortion The distortion of kerfed and unkerfed boards subjected to artificial and natural  weathering was also examined. There was large variation in the cupping, bowing and twisting of replicate board specimens and this is reflected in the large error (LSD) bars associated with the means.  Analysis of variance revealed that there were no  significant effects (p>0.05) of weathering method (artificial or natural) on the shape distortion of kerfed and unkerfed boards. Nevertheless, kerfed and unkerfed boards subjected to artificial weathering cupped more than those subjected to natural weathering. For example, the cupping of boards subjected to artificial weathering was 0.78 mm (kerfed) and 1.01 mm (unkerfed), whereas comparable figures for boards subjected to natural weathering were 0.16 mm (kerfed) and 0.38 mm (unkerfed), respectively. Figures 4.13, 4.14 and 4.15 show the effect of weathering method on cupping, bowing and twisting of kerfed and unkerfed boards, respectively.  81  p= 0.634  1.2 1.01  Cupping (mm)  1.0 0.8  lsd : 0.997  0.78  0.6 0.38  0.4 0.16  0.2 0.0  Kerfed  Unkerfed  Kerfed  Unkerfed  Natural  Artificial  Figure 4.13: The effect of weathering method on cupping of boards Figure 4.14 shows the bowing of kerfed and unkerfed boards subjected to artificial and natural weathering. Kerfed boards bowed less than unkerfed boards during artificial weathering, however, the opposite was the case for boards exposed to natural weathering. However, these differences were not statistically significant (p>0.05). p = 0.562  0.40 0.35  l.s.d :0.423  Bowing (mm)  0.30  0.254  0.25 0.20 0.15  0.142  0.126  0.106  0.10 0.05 0.00  Kerfed  Unkerfed  Artificial  Kerfed  Unkerfed  Natural  Figure 4.14: The effect of weathering method on bowing of boards 82  p = 0.186  2.0 1.8 1.6  1.41  l.s.d : 0.892  Twisting (mm)  1.4 1.2 0.91  1.0 0.8  0.75  0.6 0.31  0.4 0.2 0.0  Kerfed  Unkerfed  Artificial  Kerfed  Unkerfed  Natural  Figure 4.15: The effect of weathering method on twisting of boards The twist of boards subjected to artificial and natural weathering is shown in Figure 4.15 above. Boards exposed to natural weathering twisted more than those exposed to artificial weathering. Interestingly, the twist of unkerfed boards exposed to natural weathering was significantly (p<0.05) greater (1.41 mm) than that of boards exposed to artificial weathering (0.31 mm). In contrast, the difference in twist of kerfed boards exposed to natural and artificial weathering was not significant (p>0.05).  83  4.5 Discussion Boards exposed to natural weathering develop authentic patterns of degradation, but long testing times are required. In comparison, accelerated tests obtain results more quickly, but the pattern of degradation may differ from that which develops during natural exposure. Hence, it was very important, as mentioned in the introduction to this chapter, to determine if checking in boards exposed in the ACT resembled that in matched boards subjected to natural weathering. Results here showed that the number and sizes of checks that developed in boards subjected to natural weathering for 26 weeks or 5 days artificial weathering were not significantly different. However, the pattern of checking in boards subjected to artificial or natural weathering differed slightly. In boards subjected to artificial weathering, checks first developed at the centre of boards and these became wider and longer with time during exposure. The check in the centre of boards was invariably the largest check at the end of the cycle. However, in boards exposed to natural weathering the location of checks on the surface of boards was more random. The checking that developed in boards during natural weathering and accelerated weathering was also compared by counting the number and measuring the length of checks in scanned images of boards. The number and length of checks that developed in boards after a five day cycle in the ACT were similar to those in boards subjected to 16 or 20 weeks of natural weathering. During natural weathering, boards were removed from racks for 2 days during each week. Therefore 16 or 20 weeks of natural weathering are equivalent to 80 and 100 days of exposure, respectively. Therefore the degree to which checking was accelerated by a 5 day cycle in the accelerated check tester is ~16 to 20 times that occurring during natural exposure (in Vancouver). These figures are in the upper range (5 to 20 times) suggested by Arnold et al. (1991) for the desired acceleration of material changes during artificial weathering. The increased checking of boards exposed to artificial weathering is probably related to the fact that boards were exposed to much more frequent wetting and drying  84  and hence changes in shape than boards exposed to natural weathering. During each five day artificial weathering cycle boards were exposed to 60 dry cycles and 36 wet cycles whereas during natural exposure such cycling is more infrequent. During natural exposure we observed that boards cupped in the concave direction when they were dry and cupped in the opposite (convex) direction when they were wet. In this chapter we modified the artificial weathering cycle to induce similar changes in boards exposed in the ACT. When this was done the number of checks that developed at the surface of boards was much more than that which developed in samples subjected to Cycle 6 in Chapter 3. For example, the total length of checks that formed at the surface of the boards increased twofold and the total width of checks increased threefold. The rapid repetition of the cycling in shape of boards is probably responsible for the increased checking of boards subjected to the modified cycle in this chapter compared to the checking of boards subjected to Cycle 6 in Chapter 3. During the natural weathering trial, checks started to develop at board surfaces during the first week of exposure. The checks then gradually increased in size and number during exposure. Such early formation of checks in boards exposed to natural weathering was unexpected. One possible reason for this is that boards were subjected to a wet and dry cycle during the first five days of the trial. On the first and second day of the exposure trial 11.4 mm and 11.6 mm of rain fell, respectively. This was followed by a dry sunny period with a total of 27 hours of sunshine during the next 3 days and an average temperature and relative humidity of 10°C and 66% r.h., respectively. The early exposure of boards to this wet and dry cycle may have created stresses that were large enough to cause boards to check. There was a clear indication that most checks originated in latewood. This result accords with the findings of Coupe and Watson (1967) who suggested that during drying the stresses that develop in ray and non ray tissue are greater in latewood than in earlywood. There was some evidence that boards exposed to artificial weathering tended to cup more than boards exposed outdoors to natural weathering. This might be related to the fact that artificially weathered boards were subjected to repeated and severe wet and dry cycles in short periods of time whereas naturally weathered board were cycled  85  between such states much more slowly and infrequently. In the beginning of each artificial weathering cycle when boards were dried for 30 minutes, they cupped and became concave. The repetition of this dry cycle over short periods of time caused the boards to develop pronounced cupping, which was not fully recovered when boards were subjected to a wet cycle that induced cupping in the opposite direction. There was no indication that kerfing significantly reduced the surface checking of boards. The number of checks that developed in kerfed boards was slightly less than that in unkerfed boards, however, the difference was statistically insignificant (p>0.05). Kerfing had a slightly greater effect at reducing the cupping of boards during weathering, but again the difference was not statistically significant. These findings support the earlier findings of Ratu et al. (2007) that kerfing has little effect on the surface checking of decking boards exposed to natural weathering even though it reduced cupping of boards. Kerfs in kerfed boards became wider particularly at the lower surface of the underside of boards during the dry cycle and became narrower during the wet cycle. These observations suggested that kerfing allowed the boards to flex in response to changes in their moisture content and this may explain why it reduced the cupping of boards.  4.6 Conclusions The aim of the research in this Chapter was to examine the degree to which the Accelerated Check Tester could accelerate the development of checks at the surface of decking boards and create checking that resembled that which develops in boards during natural exposure. Boards exposed for 5 days to a modified cycle in the ACT checked to the same degree as boards exposed to 16 to 20 weeks of natural weathering and the checking that developed in artificially weathered boards was similar, but not identical to that which developed in naturally weathered boards. The width of checks was slightly wider in boards exposed in the ACT than in boards exposed to natural weathering. Furthermore, checks in boards exposed in the  86  ACT tended to develop in the centre of boards, whereas checks in boards exposed to natural weathering were more randomly distributed across the surface of boards. Kerfing slightly reduced the number of checks and distortion of boards, but in both cases the effect was not statistically significant.  87  5  Chapter Five  GENERAL CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH  5.1 General Conclusions The aim of the work presented in this thesis was to complete the development of a device (Accelerated Check Tester, ACT) and associated weathering cycle that could accelerate the surface checking of realistic-sized decking samples. The degree of checking of wood is related to the severity of moisture gradients that develop between the surface and core of wooden boards (Schniewind 1963, Stamm 1965, Evans 2004). It was hypothesized that a device that could artificially increase the severity of such moisture gradients and surface shrinkage and also their frequency of occurrence in wooden decking boards would increase surface checking. A previous study developed the first prototype of the ACT and a weathering cycle that increased the checking of wooden decking boards (Weizenegger 2006). In Chapter 3 the wetting and drying cycles employed by the ACT were altered to increase the surface drying of boards following wetting. This was achieved by increasing the temperature of drying cycles (from 73 to 80°C) or the duration of the cycle from 30 min to 60 min. However, these changes did not significantly increase the checking of wooden boards. Similarly the inclusion of a freezing step did not significantly alter the checking of boards. Therefore it can be concluded that there is no need to change the drying step in the weathering cycle that was developed for the first prototype of the ACT. Results indicated that the freezing step could be eliminated. Two other major changes were made to the weathering cycle as a result of experimental findings in Chapters 3 and 4.  88  In Chapter 3 it was found that UV exposure significantly increased checking. This finding confirmed those of Evans et al. (2008) who found that exposure to UV radiation during natural weathering increased the tendency of wood to check. Therefore, it can be concluded that a weathering cycle designed to accelerate the surface checking of wood should expose samples to UV light. A further change to the weathering cycle occurred as a result of an observation that boards exposed outdoors became convex when wet and concave when dry. The weathering cycle was modified by changing the wetting cycle to cause this to happen to samples in the ACT. This change further accelerated the development of checks on the surface of the boards. Another important factor that affected surface checking was wood species. Western red cedar was much less susceptible to checking than southern pine, and the pattern of checking in western red cedar was different from that of pine. The greater dimensional stability of western red cedar compared to southern pine may explain this finding. Kerfed and unkerfed board samples were weathered in the ACT and outdoors. Unkerfed boards tended to check more than kerfed boards, but the differences were statistically insignificant. Furthermore, kerfing had little effect on the shape distortion (cupping, bowing and twisting) of boards. Therefore, it can be concluded that kerfing has little effect on the checking and distortion of decking boards unlike the large effect that it has on the checking of poles and posts and boxed-heart square-section timber. There was clear evidence that the ACT was able to accelerate the surface checking of wooden decking boards. Results from experiments in Chapters 3 and 4 revealed that checks developed rapidly and increased in size and number over time in boards exposed in the ACT. The purpose of artificial weathering is to decrease the time consuming testing of wood, which is the major drawback of natural weathering trials. This aim of the research was successful as results from a five day weathering cycle in the ACT were comparable to those obtained after prolonged exposure of boards to natural weathering. For example, the number and total length of checks that developed were similar for boards exposed to 5 days of artificial and 20 weeks of natural weathering. During natural weathering boards were removed from racks for  89  two days during each week. Therefore 20 weeks of natural weathering equates to 100 rather than 140 days of exposure. Accordingly, results in Chapter 4 indicate that the degree of acceleration of checking in the ACT was ~20 times greater than that obtained as a result of exposure of boards to natural weathering in Vancouver. Accelerated tests also need to produce changes in the material that are similar to those observed under natural exposure conditions. In general the checking that developed in boards exposed in the ACT was similar to that which developed in boards subjected to natural weathering, except checks in boards exposed in the ACT tended to develop in the centre of boards and were slightly wider (and possibly deeper) than those in boards subjected to natural weathering. Research to correlate checking in samples exposed to artificial and accelerated weathering was restricted to check number and sizes and more subtle differences in the pattern of checking were not quantified. In addition, the results from natural weathering are valid for spring-summer exposure because the samples were weathered from May to October. Climate and other environmental conditions may strongly influence checking and it possible that different results could have been obtained if samples were exposed outdoors at different times of the year, or in different locations. The research successfully achieved its goal of developing a device and associated weathering cycle for accelerating the checking of decking boards, however, the ACT has some limitations. For instance, the ACT can only accommodate two samples at a time and this limits the speed with which different treatments and wood species can be tested. Accordingly, it would desirable to increase the size of the ACT, so that more samples can be tested simultaneously. Another limitation of the ACT is that it lacks an internal UV lamp. In this thesis samples were exposed to UV light in separate machine. It was time consuming to unfasten and refasten the samples before and after exposing them to UV light. This problem could be overcome by installing UV lights in the ACT. The research in this thesis is significant because the ACT and the associated weathering cycle is now at a point where the device is a useful tool for rapidly obtaining information on the ability of wood preservatives and water repellents to  90  restrict the checking of wood. Chemical and coating companies will benefit from using the Accelerated Check Tester to shorten the time required to test new wood protection systems.  5.2 Suggestions for Further Research The development of the ACT and associated weathering cycle provides a means of testing the susceptibility of wood to checking. There is clearly considerable scope to use the device to examine the checking of different wood species and the ability of surface treatments to reduce the severity of checking. The checking that developed in boards exposed in the ACT was similar to that in samples exposed to natural weathering. However, this research was restricted to the development of surface checking in wood. Other aspects of checking such as internal and end-checking including the depth of checks should also be investigated. Furthermore, an additional experiment to compare checking of boards subjected to natural exposure starting in fall/winter with that which develops in boards exposed in the ACT would be desirable to provide more complete information on the correlation between artificial and natural weathering. An unexpected finding was that the direction of cupping of boards influenced checking. The extent to which checks develop when boards flex and become convex compared to when they are concave is not known. It would be interesting to examine this because if checks develop when boards are convex it might be desirable to alter the weathering cycle to include a second wet cycle. As previously mentioned, the design of the ACT could be improved by increasing the number of samples that can be tested and also incorporating UV lights in the device.  91  REFERENCES Alderman D, Smith R, Araman P (2003): A profile of CCA-treated lumber removed from service in the southeastern United States decking market. Forest Products Journal 53(1), 38-45. Alexopoulus J (1992): Accelerated aging and outdoor weathering of aspen waferboard. Forest Products Journal 42(2), 15-22. Anderson E L, Pawlak Z, Owen N L, Feist W C (1991a): Infrared studies of wood weathering. Part I: Softwoods. Applied Spectroscopy 45(4), 641-647. Anderson E L, Pawlak Z, Owen N L, Feist W C (1991b): Infrared studies of wood weathering. Part II: Hardwoods. Applied Spectroscopy 45(4), 648-652. Annual Book of ASTM standards (2004a): Section 14 – General methods and instrumentation. General test methods; Forensic sciences; Terminology; Conformity assessment; Statistical methods 14(02), Easton, MD, USA. Annual Book of ASTM standards (2004b): Section 14 – General methods and instrumentation. Laboratory apparatus; Degradation of materials; Sl; Oxygen fire safety 14(04), Easton, MD, USA. Archer R R (1987): On the origin of growth stresses in trees. Wood Science and Technology 21(2), 139-154. Arndt U, Willeitner H (1969): On the resistance behaviour of wood in natural weathering. Holz als Roh- und Werkstoff 27(5), 179-188. Arnold M, Sell J, Feist W C (1991): Wood weathering in fluorescent ultraviolet and xenon arc chambers. Forest Products Journal 41(2), 40-44. Bariska M, Pizzi A, Conradie W E (1988): Structural weakening of CCA-treated timber. Holzforschung 52(5), 339-345. Belford D S, Nicholson J (1969): Emulsion additives for CCA preservatives to control weathering. Proceedings of American Wood-Preservers’ Association 65, 38-51.  92  Borgin K (1965): The effect of water repellents on the dimensional stability of wood. Norsk Skogindustri 15(11), 507-521. Borgin K (1971): The mechanism of the breakdown of the structure of wood due to environmental factors. Journal of the Institute for Wood Science 5(4), 2630. Borgin K, Corbett K (1969): Long term stabilizing effect of water repellents on wood. Timber 6(6), 5-19. Borgin K, Corbett K (1970a): The stability and weathering properties of wood treated with various oils. Plastics, Paint and Rubber 14(3), 69-72. Borgin K, Corbett K (1970b): The stability and weathering properties of wood treated with various waxes. Plastics, Paint and Rubber 14 (4), 89-94. Browne F L (1960): Wood siding left to weather naturally. Southern Lumberman 201(2513), 141-143. Browne F L, Simonson H C (1957): The penetration of light into wood. Forest Products Journal 7(10), 308-314. Chang S T, Hon D N-S, Feist W C (1982): Photodegradation and photoprotection of wood surfaces. Wood and Fiber 14(2), 104-117. Christy A G, Senden T, Evans P D (2005): Automated measurement of checks at wood surfaces. Measurement 37(2), 109-118. Cooper G A (1972): Prefreezing reduces shrinkage and alters sorption in black walnut. Forest Products Journal 22(5), 54-60. Coupe C, Watson R W (1967): Fundamental aspects of weathering. Proceedings of the Annual Convention of the British Wood Preservers’ Association, pp. 37–49. Crawford D M, DeGroot R C, Gjovik L R (1999): Ten-year performance of treated north eastern softwoods in aboveground and ground-contact exposures. USDA Forest Service Research Paper FPL-RP-578. Forest Products Laboratory, Madison, Wisconsin.  93  Cui F, Zahora A (2000): Effect of water repellent additive on the performance of ACQ treated decks. The International Research Group on Wood Preservation. Doc. No. IRG/WP/40168. Derbyshire H, Miller E R (1981): The photodegradation of wood during solar irradiation. Part I: Effects on the structural integrity of thin wood strips. Holz als Roh- und Werkstoff 39(8), 341-350. Derbyshire H, Miller E R, Turkulin H (1997): Investigations into the photodegradation of wood using microtensile testing. Part 3: The influence of temperature on photodegradation rates. Holz als Roh- und Werkstoff 55(5), 287-291. Eckelman C A (1998): The shrinking and swelling of wood and its effect on furniture. Forestry and Natural Resources. FNR 163 Evans P D (1989): Structural changes in Pinus radiata during weathering. Journal of the Institute of Wood Science 11(5), 172-181. Evans P D (2001): Wood products: weathering. In: Encyclopaedia of Materials: Science and Technology, pp 9716-9722. Eds. K.H.T. Bushow, R.W. Cahn, M.C. Flemings, B. Iischner, E.J. Kramer, S. Mahagen. Elsevier Science Ltd, Oxford. Evans P D (2008): Weathering and photoprotection of wood. In: Development of Commercial Wood Preservatives, ACS Symposium Series 982, pp 69117. Eds. T.P. Schultz, H. Militz, M.H. Freeman, B. Goodell, D.D. Nicholas. American Chemical Society, Washington. Evans P D, Donnelly C, Cunningham R B (2003): Checking of CCA-treated radiata pine decking timber exposed to natural weathering. Forest Products Journal 53(4), 66-71. Evans P D, Urban K, Chowdhury M J A (2008): Surface checking of wood is increased by photodegradation caused by ultraviolet and visible light. Wood Science and Technology 42(3), 251-265. Evans P D, Wingate-Hill R, Barry S C (2000): The effects of different kerfing and center-boring treatments on the checking of ACQ treated pine posts exposed to the weather. Forest Products Journal 50(2), 59-64.  94  Evans P D, Wingate-Hill R, Cunningham R B (1997): The ability of physical treatments to reduce checking in preservative-treated slash pine posts. Forest Products Journal 47(5), 51-55. Feist W C (1982): Weathering of wood in structural uses. In: Structural Use of Wood in Adverse Environments, pp 156-178, Eds. R.W. Meyer, R.M. Kellog. Van Nostrand Reinhold, New York. Feist W C (1983): Weathering and protection of wood. Proceedings of American Wood Preservers’ Association 79,195-205. Feist W C (1990): Outdoor wood weathering and protection. In: Archaeological Wood: Properties, Chemistry and Preservation, pp 263-298. Eds. R.M. Rowell, R.J. Barbour. American Chemical Society, Washington. Feist W C, Hon D N-S (1984): Chemistry of weathering and protection. In: Chemistry of Solid Wood, pp. 401-451. Ed. R.M. Rowell. American Chemical Society, Washington. Feist W C, Mraz E A (1978): Comparison of outdoor and accelerated weathering of unprotected softwoods. Forest Products Journal 28(3), 38-43. Feist W C, Williams S R (1991): Weathering durability of chromium-treated southern pine. Forest Products Journal 41(1), 8-14. Flæte P O, Høibø O A, Fjærtoft F, Nilsen T-N (2000): Crack formation in unfinished siding of aspen (Populus tremula L.) and Norway spruce (Picea abies (L.) Karst.) during accelerated weathering. Holz als Roh- und Werkstoff 58(3), 135-139. Fredonia Group Inc. (2005): Wood and competitive decking. [Online; cited 16th August 2006; 14:00 PST]. Available from: MarketResearch.com url: http://www.marketresearch.com/product/display.asp?productid=1175959 Furuno T (2001): Histochemical study of wood deterioration by UV irradiation and prevention of deterioration using poly PEGMA. In: High-Performance Utilization of Wood for Outdoor Uses. Ed. Y. Imamura. Wood Research Institute, Kyoto University, Kyoto, Japan.  95  Futó L P (1974): Der photochemische Abbau des Holzes als Präparations- und Analysemethode. Holz als Roh- und Werkstoff 32(8), 303-314. Gilfedder J, Keating W G, Robertson I (1968): Influence of certain preservatives on pole splitting. Forest Products Journal 18(1), 28-30. Gjelsvik T (1983): Apparatus for accelerated weathering of building materials and components. Materials and Structures 16(3), 209-211. Graham R D (1973): Preventing and stopping internal decay of Douglas fir poles. Holzforschung 27(5), 168-173. Graham R D, Estep E M (1966): Effect of incising and saw kerfs on checking of pressure treated Douglas fir spar crossarms. Proceedings of American Wood-Preservers’ Association 62, 155-160. Goodell B (1991): Preservative treatment and field test monitoring of spruce pole stock; pressure and diffusible chemical treatments. The International Research Group on Wood Preservation. Doc. No. IRG/WP/3605. Groves K W, Banana A Y (1986): Weathering characteristics of Australian grown radiata pine. Journal of the Institute of Wood Science 10(5), 210-213. Helsing G, Graham R D (1976): Saw kerfs reduce checking and prevent internal decay in pressure-treated Douglas-fir poles. Holzforschung 30(6), 184186. Henry W T (1970): Effect of incising on the reduction of serious checking in incised vs. unincised hardwood crossties. Proceedings of American WoodPreservers’ Association 66, 171-178. Hon D N-S (1981): Photochemical degradation of lignocellulosic materials. In: Developments in Polymer Degradation, pp 229 – 281. Ed. N. Grassie. Applied Science Publishers, London. Hon D N-S (1983): Weathering reactions and protection of wood surfaces. Journal of Applied Polymer Science, Applied Polymer Symposium 37(1), 845864.  96  Hon D N-S, Chang S-T (1984): Surface degradation of wood by ultraviolet light. Journal of Polymer Science, Polymer Chemistry Edition, 22(9), 22272241. Hon D N-S, Chang S-T, Feist W C (1985): Protection of wood surfaces against photooxidation. Journal of Applied Polymer Science 20(4), 1429-1448. Hon D N-S, Ifju G (1978): Measuring penetration of light into wood by detection of photo-induced free radicals. Wood Science 11(2), 118-127. Ibach, RE (1999): Wood preservation. USDA Forest Service Research Paper FPLGTR-113. Forest Products Laboratory, Madison, Wisconsin. Imamura Y (2001): Preface. In: High-Performance Utilization of Wood for Outdoor Uses, pp i-iii. Ed. Y. Imamura. Wood Research Institute, Kyoto University, Kyoto, Japan. Jin L, Archer K, Preston A (1991): Surface characteristics of wood treated with various AAC, ACQ and CCA formulations after weathering. The International Research Group on Wood Preservation. Doc. No. IRG/WP/2369. Kabir F R A, Nicholas D D, Vasishth R, Barnes M (1992): Laboratory methods to predict the weathering characteristics of wood. Holzforschung 46(5), 395401. Kamden D P, Zhang J (2000): Characterization of checks and cracks on the surface of weathered wood. The International Research Group on Wood Preservation. Doc. No. IRG/WP/40153. Kataoka Y, Kiguchi M, Evans P D (2004): Photodegradation depth profile and penetration of light in Japanese cedar earlywood (Cryptomeria japonica D. Don) exposed to artificial solar radiation. Surface Coatings International Part B: Coatings Transactions 87(B3), 187-193. Kishino M, Nakano T (2004a): Artificial weathering of tropical wood. Part 1: Changes in wettability. Holzforschung 58(5), 552-557. Kishino M, Nakano T (2004b): Artificial weathering of tropical wood. Part 2: Color change. Holzforschung 58(5), 558-565.  97  Kurisaki H (2004): Effect of kerfing on preventing the surface check and reducing inner decay risk in sugi post. Proceedings of the 3rd International Symposium on Surfacing and Finishing of Wood, 336-341. Kühne H, Hochweber M, Sell J (1968): Freiland-Bewitterungsversuche an Außenanstrichen für Holz. Versuchszeitraum 1962-1967. Report No. 82. Eidgenössische Materialprüfungs- und Versuchsanstalt für Industrie, Bauwesen und Gewerbe, Dübendorf. Levi M P, Coupe C, Nicholson J (1970). Distribution and effectiveness in Pinus sp. of a water-repellent additive for water-borne wood preservatives. Forest Products Journal 20(11), 32-37. Mackay J F G (1973): Surface checking and drying behaviour of Pinus radiata sapwood boards treated with CCA preservative. Forest Products Journal 23(9), 92-97. Mater M H (1972): Boring 40-foot long utility poles for conduit passage. Forest Products Journal 22(6), 28-31. McIntosh D C (1955): Shrinkage of red oak and beech. Forest Products Journal 5(2), 355-359. McQueen J, Stevens J (1998): Disposal of CCA-treated wood. Forest Products Journal 48(11/12), 86-90. Miller E R, Derbyshire H (1981): The photodegradation of wood during solar irradiation. In: Proceedings of the second international conference on the durability of building materials and components. September 14-16, 1981, pp 279-287. National Bureau of Standards, Gaithersburg, Maryland. Miniutti V P (1967): Microscopic observations of ultraviolet irradiated and weathered softwood surfaces and clear coatings. USDA Forest Service Research Paper FPL 74. Forest Products Laboratory, Madison, Wisconsin. Morrell J J (1990): Effect of kerfing on performance of Douglas-fir utility poles in the Pacific Northwest. The International Research Group on Wood Preservation. Document No: IRG/WP/3604. Morrell J J, Newbill M A (1986): Kerfing to prevent decay of Douglas fir poles: an update. Forest Products Journal 36(5), 46-48.  98  Newbill M A (1997): Through-boring improves preservatives treatment and extends service life. Wood Pole Newsletter 21, 3-5. Owen J A, Owen N L, Feist W C (1993): Scanning electron microscope and infrared studies of weathering in Southern pine. Journal of Molecular Structure 300(1-2), 105-114. Placket D V, Chittenden C M, Preston A F (1984): Exterior weathering trials on Pinus radiata roofing shingles. New Zealand Journal of Forestry Science 14(3), 368-81. Placket D V, Dunningham E A, Singh A P (1992): Weathering of chemically modified wood; accelerated weathering of acetylated radiata pine. Holz als Roh- und Werkstoff 50(4), 135-140. Principia Partners (2002): Polymer-wood composites will double share of giant market for boards and railings in residential decking by 2005. [Online; cited 22nd September 2006; 10:00 PST]. Available from: http://www.principiaconsulting.com/ Ratu R N, Weizenegger J, Evans P D (2007): Preliminary observations of the effect of kerfing on the surface checking and warping of flat sawn southern pine decking. The International Research Group on Wood Protection. Doc. No. IRG/WP 07-20360. Ratu R N, Evans, P D (2008): Development of a weatherometer to accelerate the surface checking of wood. The International Research Group on Wood Protection. Doc. No. IRG/WP 08-20388 Rhatigan R G, Morrell J J (2003): Use of through-boring to improve CCA or ACZA treatment of refractory Douglas-fir and grand fir. Forest Products Journal 53(6):33-35 Rowell R M (1999): Wood handbook—Wood as an engineering material. USDA General Technical FPL–GTR–113. Forest Products Laboratory, Madison, Wisconsin. Rowell R M, Feist W C, Ellis W D (1981): Weathering of chemically modified southern pine. Wood Science 13(4), 202-208.  99  Ruddick J N R (1981): The effect of kerfing on check formation in treated white spruce (Picea glauca) poles. The International Research Group on Wood Preservation. Doc. No. IRG/WP/3167. Ruddick J N R (1988): Kerfing reduces checking in ACA-treated Western white spruce poles. The International Research Group on Wood Preservation. Doc. No. IRG/WP/3477. Ruddick J N R, Ross N A (1979): Effect of kerfing on checking of untreated Douglas-fir pole sections. Forest Products Journal 29(9), 27-30. Shah V (1983): Handbook of plastic testing technology. John Wiley and Sons, New York. Schoeman M, Dickinson, D (1997): Growth of Aureobasidium pullulans on lignin breakdown products at weathered wood surfaces. Mycologist 11(4),168172. Sadoh T (2001): Claim Ga Ohi Curui – Doh Shite Fugesu (in Japanese). In : Sadoh T “Ki Ga Wakaru”, Gakugei Shuppan-Sha, Kyoto, 116-119. Saitoh S, Nakajima Y, Gensai H, Iwatani I (2000): Kiln drying test of full scale sugi (Cryptomeria japonica) boxed heart timber (in Japanese), Transaction of the Annual Meeting of Kanto branch of the Japanese Forestry Society, 52, 173-176 Sandberg D (1996): The influence of pith and juvenile wood on proportion of cracks in sawn timber when kiln dried and exposed to wetting cycles. Holz als Roh- und Werstoff 54(3), 152. Sandberg D (1997): Radially sawn timber. The influence of annual ring orientation on crack formation and deformation in water soaked pine (Pinus sylvestris L.) and spruce (Picea abies (L.) Kast.) timber. Holz als Roh- und Werkstoff 55(3), 175-182. Sandberg D (1999): Weathering of radial and tangential wood surfaces of pine and spruce. Holzforschung 53(4), 355-354. Sandberg D (2005): Distortion and visible crack formation in green and seasoned timber: influence of annual ring orientation in the cross section. Holz als Roh- und Werkstoff 63(1), 11-18.  100  Sandberg D, Söderström O (2006): Crack formation due to weathering of radial and tangential sections of pine and spruce. Wood Material Science and Engineering 1(1), 12-20. Schniewind A P (1959): Transverse anisotropy of wood: a function of gross anatomic structure. Forest Products Journal 9(10), 350-59. Schniewind A P (1963): Mechanism of check formation. Forest Products Journal 13(11), 475-480. Searle S R, Casella G, McCulloch C E (1992): Variance components analysis. Wiley Interscience. New York, 501 pp. Sell J (1968): Investigations on the infestation of untreated and surface treated wood by blue-stain fungi. Holz als Roh- und Werkstoff 26(6), 215-222. Sell J, Leukens U (1969): Über die außenklimatische und biologische Beanspruchung von unbehandelten und angestrichenen Holzoberflächen. Oberfläche – Surface 10(8), 536-539. Sell J, Leukens U (1971a). Untersuchungen an bewitterten Holzoberflächen – Zweite Mitteilung: Verwitterungserscheinungen an ungeschützten Hölzern. Holz als Roh- und Werkstoff 39(1), 23-31. Sell J, Wälchli O (1969): Changes in the surface texture of weather-exposed wood. Material und Organismen 4(2), 81-87. Stamm A J (1964): Wood and Cellulose Science. The Ronald Press Co, New York. Stamm A J (1965): Modification of wood for improved finishing. Official Digest. Journal of Paint Technology 37, 707-719. Stamm A J, Loughborough W K (1942): Variations in shrinking and swelling of wood. Transactions of the American Society for Mechanical Engineering 64, 379-285. Suits L D, Hsuan Y G (2003): Assessing the photo-degradation of geosynthetics by outdoor exposure and laboratory weatherometer. Geotextile and Geomembranes 21(3), 111-122.  101  Temiz A, Yildiz U C, Aydin I, Eikenes M, Alfredsen G, Colakoglu G (2005): Surface roughness and color characteristic of wood treated with preservatives after accelerated weathering test. Applied Surface Science 250(1-4), 35-42. Temiz A, Terziev N, Eikenes M, Hafren J (2007): Effect of accelerated weathering on surface chemistry of modified wood. Applied Surface Science 253(12), 5355-5362. Turkulin H, Sell J (2002): Investigations into the photodegradation of wood using microtensile testing. Part 4: Tensile properties and fractography of weathered wood. Holz als Roh- und Werkstoff 60(2), 69-105. Urban K, Evans P D (2005): Preliminary observations of the effect of growth ring orientation on the surface checking of flat sawn southern pine decking. The International Research Group on Wood Protection. Doc. No. IRG/WP 05-50. Webb D A, Sullivan J D (1964): Surface effect of light and water on wood. Forest Products Journal 14(11), 531-534. Weizenegger J (2006): Accelerated Surface Checking of Southern Pine. Diploma Thesis. Berner Fachhoschule, University of Applied Sciences HSB (Hochschule Bern). Wengert E M (1966): Parameters for predicting maximum surface temperature of wood in exterior exposure. USDA Forest Service Research Paper FPL 62. Forest Products Laboratory, Madison, Wisconsin. Williams R S (1987): Acid effects on accelerated wood weathering. Forest Products Journal 37(2), 37-38. Williams R S, Feist W C (1999): Water repellents and water-repellent preservatives for wood. Gen. Tech. Rep. FPL-GTR-109. U.S Dept. of Agriculture, Forest Service, 12p Yata S (2001): Occurrence of drying checks in softwood during outdoor exposure. In: High-Performance Utilization of Wood for Outdoor Uses, pp 65-70 Ed. Y. Imamura. Wood Research Institute, Kyoto University, Kyoto, Japan.  102  Zahora A R (1991): Interactions between water-borne preservatives and emulsion additives that influence the water repellency of wood. The International Research Group of Wood Preservation. Doc. No. IRG/WP/2374. Zahora A R (1992): A water repellent additive’s influence on field performance of southern yellow pine lumber. American Wood-Preservers’ Association, 148-159. Zahora A R (2000): Long-term performance of a wax type additive for use with water-borne pressure preservative treatments. The International Research Group on Wood Preservation. Doc. No. IRG/WP/40159.  103  APPENDICES APPENDIX 1 Results for 1st experiment interpreted in graphs as described in Chapter 3 ________________________________________ GenStat Fifth Edition (Service Pack 1) GenStat Procedure Library Release PL12.1 ________________________________________ 1 %CD 'C:/Documents and Settings/phil.evans/Desktop' 2 "Data taken from File: \ -3 C:/Documents and Settings/phevans/My Documents/Postgrads/Ricky Ratu/Copy of Exp #1 Check.xls\ -4 " 5 DELETE [Redefine=yes] _stitle_: TEXT _stitle_ 6 READ [print=*;SETNVALUES=yes] _stitle_ 10 PRINT [IPrint=*] _stitle_; Just=Left Data imported from Excel file: C:\Documents and Settings\phevans\My Documents\Postgrads\Ricky Ratu\Copy of Exp #1 Check.xls on: 8-Sep-2007 10:01:37 taken from sheet "Check", cells A1:H49 11 DELETE [redefine=yes] Block,Run,Board,Cycle,Species,No,Length,Width 12 FACTOR [modify=yes;nvalues=48;levels=4] Block 13 READ Block; frepresentation=ordinal Identifier Block 16 17  Missing 0  Levels 6  Values 48  Missing 0  Levels 2  Values 48  Missing 0  Levels 6  FACTOR [modify=yes;nvalues=48;levels=2;labels=!t('Syp','Wrc')] Species READ Species; frepresentation=ordinal  Identifier Species 32 33  Values 48  FACTOR [modify=yes;nvalues=48;levels=6] Cycle READ Cycle; frepresentation=ordinal  Identifier Cycle 28 29  Levels 4  FACTOR [modify=yes;nvalues=48;levels=2] Board READ Board; frepresentation=ordinal  Identifier Board 24 25  Missing 0  FACTOR [modify=yes;nvalues=48;levels=6] Run READ Run; frepresentation=ordinal  Identifier Run 20 21  Values 48  Values 48  Missing 0  Levels 2  VARIATE [nvalues=48] No READ No  Identifier No  Minimum 3.000  Mean 27.06  Maximum 97.00  Values 48  Missing 0  104  36 37  VARIATE [nvalues=48] Length READ Length  Identifier Length 41 42  Minimum 37.00  Mean 527.1  Maximum 2027  Values 48  Missing 0  Maximum 5.000  Values 48  Missing 0  Skew  VARIATE [nvalues=48] Width READ Width  Identifier Width  Minimum 0.1500  Mean 1.872  46 RESTRICT Block,Run,Board,Cycle,Species,No,Length,Width 48 "Split-Plot Design." 49 BLOCK Block/Run/Board 50 TREATMENTS Cycle*Species 51 COVARIATE "No Covariate" 52 ANOVA [PRINT=aovtable,information,means; FACT=32; FPROB=yes; PSE=diff,lsd; LSDLEVEL=5]\  105  ***** Analysis of variance ***** Variate: No Source of variation Block stratum  d.f. 3  s.s. 274.6  m.s. 91.5  v.r. F pr. 0.55  Block.Run stratum Cycle Residual  5 15  3124.9 2479.8  625.0 165.3  3.78 0.73  0.020  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  11071.7 1299.4 4058.4  11071.7 259.9 225.5  49.11 1.15  <.001 0.370  Total  47  22308.8  * MESSAGE: the following units have large residuals. Block 1  Run 6  Block 2 Block 2  Run 6 Run 6  15.1  s.e. 7.2  Board 1 Board 2  -22.3 22.3  s.e. 9.2 s.e. 9.2  4 26.4  5 28.0  ***** Tables of means ***** Variate: No Grand mean  27.1  Cycle  1 24.3  2 21.3  Species  Syp 42.2  Wrc 11.9  Cycle 1 2 3 4 5 6  Species  Syp 39.5 33.2 26.3 41.5 44.5 68.5  3 18.8  6 43.7  Wrc 9.0 9.3 11.3 11.3 11.5 19.0  *** Standard errors of differences of means *** Table  Cycle  Species  Cycle Species rep. 8 24 4 s.e.d. 6.43 4.33 9.88 d.f. 15 18 32.87 Except when comparing means with the same level(s) of Cycle 10.62 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 13.70 9.11 20.11 d.f. 15 18 32.87 Except when comparing means with the same level(s) of Cycle 22.31 d.f. 18  106  Variate: Width Source of variation Block stratum  d.f. 3  s.s. 3.2214  m.s. 1.0738  v.r. F pr. 2.82  Block.Run stratum Cycle Residual  5 15  5.0073 5.7195  1.0015 0.3813  2.63 0.53  0.067  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  53.4463 3.9715 12.8734  53.4463 0.7943 0.7152  74.73 1.11  <.001 0.389  Total  47  84.2395  * MESSAGE: the following units have large residuals. Block 2 Block 2  Run 6 Run 6  Board 1 Board 2  -1.10 1.10  s.e. 0.52 s.e. 0.52  4 1.51  5 1.81  ***** Tables of means ***** Variate: Width Grand mean  1.87  Cycle  1 1.91  2 1.80  Species  Syp 2.93  Wrc 0.82  Cycle 1 2 3 4 5 6  Species  Syp 3.21 2.90 2.29 2.43 2.70 4.04  3 1.66  6 2.54  Wrc 0.60 0.70 1.04 0.60 0.93 1.04  *** Standard errors of differences of means *** Table  Cycle  Species  Cycle Species rep. 8 24 4 s.e.d. 0.309 0.244 0.524 d.f. 15 18 31.55 Except when comparing means with the same level(s) of Cycle 0.598 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle Species rep. 8 24 4 l.s.d. 0.658 0.513 1.067 d.f. 15 18 31.55 Except when comparing means with the same level(s) of Cycle 1.256 d.f. 18 71  Cycle  Species  DAPLOT fitted,normal,halfnormal,histogram  107  ***** Analysis of variance ***** Variate: Length Source of variation Block stratum  d.f. 3  s.s. 124775.  m.s. 41592.  v.r. 0.76  F pr.  Block.Run stratum Cycle Residual  5 15  1468848. 824956.  293770. 54997.  5.34 0.86  0.005  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  5702855. 558457. 1156748.  5702855. 111691. 64264.  88.74 1.74  <.001 0.177  Total  47  9836638.  * MESSAGE: the following units have large residuals. Block 1  Run 6  Block 2 Block 2  Run 6 Run 6  326.  s.e. 131.  Board 1 Board 2  -374. 374.  s.e. 155. s.e. 155.  4 452.  5 563.  ***** Tables of means ***** Variate: Length Grand mean  527.  Cycle  1 439.  2 429.  Species  Syp 872.  Wrc 182.  Cycle 1 2 3 4 5 6  Species  Syp 757. 743. 577. 752. 956. 1447.  3 381.  6 899.  Wrc 121. 115. 186. 151. 171. 350.  *** Standard errors of differences of means *** Table  Cycle  Species  Cycle Species rep. 8 24 4 s.e.d. 117.3 73.2 172.7 d.f. 15 18 33 Except when comparing means with the same level(s) of Cycle 179.3 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 249.9 153.7 351.3 d.f. 15 18 33 Except when comparing means with the same level(s) of Cycle 376.6 d.f. 18  108  ***** Analysis of variance ***** Variate: Shape Source of variation Block stratum  d.f. 3  s.s. 2.746E+07  m.s. 9.154E+06  v.r. 0.33  F pr.  Block.Run stratum Cycle Residual  5 15  6.621E+08 4.110E+08  1.324E+08 2.740E+07  4.83 1.17  0.008  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  1.435E+09 2.334E+08 4.229E+08  1.435E+09 4.669E+07 2.349E+07  61.07 1.99  <.001 0.129  Total  47  3.192E+09  * MESSAGE: the following units have large residuals. Block 1  Run 6  Block 2 Block 2  Run 6 Run 6  8052.  s.e. 2926.  Board 1 Board 2  -7730. 7730.  s.e. 2968. s.e. 2968.  ***** Tables of means ***** Variate: Shape Grand mean  8361.  Cycle  1 6403.  2 5835.  Species  Syp 13828.  Wrc 2894.  Cycle 1 2 3 4 5 6  Species  Syp 10809. 9932. 7517. 13314. 15715. 25680.  3 4784.  4 7998.  5 9097.  6 16048.  Wrc 1997. 1737. 2051. 2683. 2479. 6415.  *** Standard errors of differences of means *** Table  Cycle  Species  Cycle Species rep. 8 24 4 s.e.d. 2617.3 1399.2 3567.0 d.f. 15 18 32.09 Except when comparing means with the same level(s) of Cycle 3427.3 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 5578.6 2939.6 7264.9 d.f. 15 18 32.09 Except when comparing means with the same level(s) of Cycle 7200.6 d.f. 18  109  GenStat Release 4.21 (PC/Windows XP) 08 September 2007 10:14:53 Copyright 2001, Lawes Agricultural Trust (Rothamsted Experimental Station) ________________________________________ GenStat Fifth Edition (Service Pack 1) GenStat Procedure Library Release PL12.1 ________________________________________ 1  %CD 'C:/Documents and Settings/phil.evans/Desktop'  2 "Data taken from File: \ -3 C:/Documents and Settings/phevans/My Documents/Postgrads/Ricky Ratu/Copy of Exp #1 Check.xls\ -4 " 5 DELETE [Redefine=yes] _stitle_: TEXT _stitle_ 6 READ [print=*;SETNVALUES=yes] _stitle_ 10 PRINT [IPrint=*] _stitle_; Just=Left Data imported from Excel file: C:\Documents and Settings\phevans\My Documents\P ostgrads\Ricky Ratu\Copy of Exp #1 Check.xls on: 8-Sep-2007 10:15:34 taken from sheet "Shape distortion", cells A1:M49 11 12 13 14  DELETE [redefine=yes] Block,Run,Board,Cycle,Species,Cup_A,Cup_B,Cup_C,\ Twist_A,Twist_B,Bow_A,Bow_B,Bow_S FACTOR [modify=yes;nvalues=48;levels=4] Block READ Block; frepresentation=ordinal  Identifier Block 17 18  Levels 6  Values 48  Missing 0  Levels 2  Values 48  Missing 0  Levels 6  Values 48  Missing 0  Levels 2  VARIATE [nvalues=48] Cup_A READ Cup_A  Identifier Cup_A 39 40  Missing 0  FACTOR [modify=yes;nvalues=48;levels=2;labels=!t('Syp','Wrc')] Species READ Species; frepresentation=ordinal  Identifier Species 33 34  Values 48  FACTOR [modify=yes;nvalues=48;levels=6] Cycle READ Cycle; frepresentation=ordinal  Identifier Cycle 29 30  Levels 4  FACTOR [modify=yes;nvalues=48;levels=2] Board READ Board; frepresentation=ordinal  Identifier Board 25 26  Missing 0  FACTOR [modify=yes;nvalues=48;levels=6] Run READ Run; frepresentation=ordinal  Identifier Run 21 22  Values 48  Minimum 0.1400  Mean 0.7587  Maximum 1.830  Values 48  Missing 0  VARIATE [nvalues=48] Cup_B READ Cup_B  110  Identifier Cup_B 45 46  Minimum 0.1700  Mean 0.9694  Maximum 2.420  Values 48  Missing 0  Minimum -2.380  Mean -0.1398  Maximum 2.270  Values 48  Missing 0  Minimum Mean -0.9600 -0.63E-03  Maximum 2.140  Values 48  Missing 0  Minimum 0.01000  Mean 0.5700  Maximum 2.140  Values 48  Missing 0  Maximum 1.950  Values 48  Missing 0  Maximum 4.890  Values 48  Missing 0  VARIATE [nvalues=48] Bow_B READ Bow_B  Identifier Bow_B 75 76  Missing 0  VARIATE [nvalues=48] Bow_A READ Bow_A  Identifier Bow_A 69 70  Values 48  VARIATE [nvalues=48] Twist_B READ Twist_B  Identifier Twist_B 63 64  Maximum 2.960  VARIATE [nvalues=48] Twist_A READ Twist_A  Identifier Twist_A 57 58  Mean 0.8575  VARIATE [nvalues=48] Cup_C READ Cup_C  Identifier Cup_C 51 52  Minimum 0.1300  Minimum -0.1100  Mean 0.5452  VARIATE [nvalues=48] Bow_S READ Bow_S  Identifier Bow_S  Minimum 0.01000  Mean 0.8631  Skew  81 RESTRICT Block,Run,Board,Cycle,Species,Cup_A,Cup_B,Cup_C,Twist_A,Twist_B,\ 82 Bow_A,Bow_B,Bow_S 83 84 "Split-Plot Design." 85 BLOCK Block/Run/Board 86 TREATMENTS Cycle*Species 87 COVARIATE "No Covariate" 88 ANOVA [PRINT=aovtable,information,means; FACT=32; FPROB=yes; PSE=diff,lsd; LSDLEVEL=5]\ 89 Cup_A  111  ***** Analysis of variance ***** Variate: Cup_C Source of variation Block stratum  d.f. 3  s.s. 1.5495  m.s. 0.5165  v.r. 2.31  F pr.  Block.Run stratum Cycle Residual  5 15  1.2672 3.3590  0.2534 0.2239  1.13 1.97  0.386  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  5.2735 0.8174 2.0488  5.2735 0.1635 0.1138  46.33 1.44  <.001 0.259  Total  47  14.3155  * MESSAGE: the following units have large residuals. Block 1  Run 5  Block Block Block Block  Run Run Run Run  1 1 1 1  4 4 6 6  0.684 Board Board Board Board  s.e. 0.265  1 2 1 2  0.444 -0.444 0.479 -0.479  s.e. s.e. s.e. s.e.  0.207 0.207 0.207 0.207  ***** Tables of means ***** Variate: Cup_C Grand mean  0.969  Cycle  1 1.034  2 0.810  Species  Syp 1.301  Wrc 0.638  Cycle 1 2 3 4 5 6  Species  Syp 1.500 1.102 0.858 1.530 1.255 1.560  3 0.685  4 1.094  5 1.095  6 1.099  Wrc 0.567 0.517 0.513 0.657 0.935 0.637  *** Standard errors of differences of means *** Table  Cycle  Species  Cycle Species rep. 8 24 4 s.e.d. 0.2366 0.0974 0.2906 d.f. 15 18 28.08 Except when comparing means with the same level(s) of Cycle 0.2386 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 0.5043 0.2046 0.5952 d.f. 15 18 28.08 Except when comparing means with the same level(s) of Cycle 0.5012 d.f. 18  112  ***** Analysis of variance ***** Variate: Bow_S Source of variation Block stratum  d.f. 3  s.s. 8.1896  m.s. 2.7299  v.r. 4.71  F pr.  Block.Run stratum Cycle Residual  5 15  3.0179 8.6988  0.6036 0.5799  1.04 0.82  0.430  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  6.7425 2.1421 12.7106  6.7425 0.4284 0.7061  9.55 0.61  0.006 0.696  Total  47  41.5014  * MESSAGE: the following units have large residuals. Block 1  Run 6  Block Block Block Block  Run Run Run Run  1 1 1 1  4 4 6 6  1.00 Board Board Board Board  s.e. 0.43  1 2 1 2  1.13 -1.13 1.52 -1.52  s.e. s.e. s.e. s.e.  0.51 0.51 0.51 0.51  ***** Tables of means ***** Variate: Bow_S Grand mean  0.86  Cycle  1 0.61  2 0.61  Species  Syp 1.24  Wrc 0.49  Cycle 1 2 3 4 5 6  Species  Syp 0.76 0.74 0.90 1.75 1.43 1.85  3 0.66  4 1.13  5 0.95  6 1.22  Wrc 0.47 0.48 0.42 0.51 0.47 0.59  *** Standard errors of differences of means *** Table Cycle Species Cycle Species rep. 8 24 4 s.e.d. 0.381 0.243 0.567 d.f. 15 18 33 Except when comparing means with the same level(s) of Cycle 0.594 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 0.812 0.510 1.154 d.f. 15 18 33 Except when comparing means with the same level(s) of Cycle 1.248 d.f. 18  113  ***** Analysis of variance ***** Variate: Twist_A Source of variation Block stratum  d.f. 3  s.s. 4.2828  m.s. 1.4276  v.r. 1.43  F pr.  Block.Run stratum Cycle Residual  5 15  2.8042 14.9861  0.5608 0.9991  0.56 1.32  0.728  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  0.0463 2.0020 13.5779  0.0463 0.4004 0.7543  0.06 0.53  0.807 0.750  Total  47  37.6993  * MESSAGE: the following units have large residuals. Block 1 Run 2 1.17 s.e. 0.56 Block Block Block Block  1 1 2 2  Run Run Run Run  4 4 6 6  Board Board Board Board  1 2 1 2  1.20 -1.20 1.30 -1.30  s.e. s.e. s.e. s.e.  0.53 0.53 0.53 0.53  ***** Tables of means ***** Variate: Twist_A Grand mean  -0.14  Cycle  1 0.02  2 -0.19  Species  Syp -0.17  Wrc -0.11  Cycle 1 2 3 4 5 6  Species  Syp -0.32 -0.31 -0.50 0.52 0.04 -0.46  3 -0.46  4 0.30  5 -0.20  6 -0.30  Wrc 0.35 -0.07 -0.43 0.07 -0.43 -0.14  *** Standard errors of differences of means *** Table Cycle Species Cycle Species rep. 8 24 4 s.e.d. 0.500 0.251 0.662 d.f. 15 18 31.32 Except when comparing means with the same level(s) of Cycle 0.614 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 1.065 0.527 1.350 d.f. 15 18 31.32 Except when comparing means with the same level(s) of Cycle 1.290 d.f. 18  114  ***** Analysis of variance ***** Variate: Twist_B Source of variation Block stratum  d.f. 3  s.s. 1.7733  m.s. 0.5911  v.r. 1.23  F pr.  Block.Run stratum Cycle Residual  5 15  2.5087 7.2083  0.5017 0.4806  1.04 0.96  0.428  Block.Run.Board stratum Species Cycle.Species Residual  1 5 18  0.2930 0.6588 9.0434  0.2930 0.1318 0.5024  0.58 0.26  0.455 0.928  Total  47  21.4855  * MESSAGE: the following units have large residuals. Block Block Block Block  1 1 2 2  Run Run Run Run  6 6 6 6  Board Board Board Board  1 2 1 2  -1.42 1.42 0.95 -0.96  s.e. s.e. s.e. s.e.  0.43 0.43 0.43 0.43  ***** Tables of means ***** Variate: Twist_B Grand mean  0.00  Cycle  1 0.14  2 -0.10  Species  Syp -0.08  Wrc 0.08  Cycle 1 2 3 4 5 6  Species  Syp -0.09 -0.21 -0.30 -0.05 -0.13 0.32  3 -0.17  4 -0.20  5 -0.12  6 0.45  Wrc 0.36 0.02 -0.04 -0.34 -0.11 0.58  *** Standard errors of differences of means *** Table Cycle Species Cycle Species rep. 8 24 4 s.e.d. 0.347 0.205 0.496 d.f. 15 18 32.84 Except when comparing means with the same level(s) of Cycle 0.501 d.f. 18 *** Least significant differences of means (5% level) *** Table  Cycle  Species  Cycle Species rep. 8 24 4 l.s.d. 0.739 0.430 1.009 d.f. 15 18 32.84 Except when comparing means with the same level(s) of Cycle 1.053 d.f. 18  115  Data from 1st Experiment (Chapter 3) (for Appendix 1) Block  Run  Board  Cycle  Species  1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4  1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6  1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2  1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6  Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc  Number 24 13 20 16 13 11 70 7 59 15 97 29 32 14 33 7 20 26 27 13 43 23 35 30 36 4 39 11 25 4 35 21 37 5 79 12 66 5 41 3 47 4 34 4 39 3 63 5  Total Check Length (mm) 490 167 658 204 313 257 1301 183 965 201 2027 595 822 191 695 106 333 392 519 147 928 385 945 595 746 62 1004 112 656 47 652 231 745 57 1309 150 970 63 614 37 1004 50 536 44 1184 42 1507 60  Width (mm) 3.2 1.3 3.1 1.8 1.55 1.3 4 0.5 3.2 0.8 5 1.6 2.95 0.65 2.95 0.3 2.05 2.45 1.45 0.65 2.45 2.5 2.4 1.6 3.35 0.2 3.4 0.55 2.7 0.2 2.55 1.05 2.2 0.25 4.15 0.7 3.35 0.25 2.15 0.15 2.85 0.2 1.7 0.2 2.95 0.15 4.6 0.25  L/W ratio 3,685 1,670 4,427 1,850 2,695 2,350 23,120 2,290 17,580 3,820 39,630 10,840 10,655 3,820 9,765 2,120 3,855 3,915 9,720 2,940 16,960 4,114 14,645 10,840 9,708 1,240 14,165 2,240 7,100 940 9,697 4,620 12,560 1,140 24,775 2,780 19,190 1,260 11,370 740 16,420 1,000 10,720 880 15,762 840 23,672 1,200  116  Block  Run  Board  Cycle  Species  1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4  1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6  1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2  1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6 1 1 2 2 3 3 4 4 5 5 6 6  Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc Sp Wrc  Shape Distortion (mm) Cup C Twist A Bow S 1.44 -1.51 0.41 0.38 -0.23 0.52 0.67 1.05 0.25 0.17 0.18 0.48 0.69 -1.1 1.02 0.37 -0.3 0.3 2.16 2.27 3.73 0.4 -0.57 0.23 2.01 -2.26 3.47 1.78 -0.82 0.65 2.42 -2.38 4.89 0.54 -0.43 0.58 2.26 1.33 0.79 0.73 1.53 0.54 1.38 -0.32 1.13 0.76 0.09 0.61 1.43 -0.49 1.13 0.74 -0.78 0.52 1.71 0.27 1.04 0.6 0.33 0.72 1.12 0.93 1.34 0.89 -0.36 0.65 1.76 1.68 1.21 1.03 -0.61 0.72 1.14 -0.37 0.5 0.49 -0.28 0.01 1.57 -1.34 0.64 0.72 -0.6 0.13 0.73 -0.72 0.61 0.34 -0.41 0.21 0.99 0.26 0.09 0.51 0.37 0.04 1.26 0.38 0.2 0.51 -0.22 0.1 1.02 -0.73 0.18 0.59 0.21 0.02 1.16 -0.72 1.34 0.67 0.38 0.8 0.79 -0.62 0.93 0.42 0.03 0.7 0.58 0.32 0.82 0.6 -0.22 0.63 1.26 -0.72 2.13 1.12 0.16 1.06 0.63 1.1 0.73 0.56 -0.32 0.46 1.04 -0.41 1.13 0.39 0.26 1.04  117  APPENDIX 2 Results for 2nd experiment interpreted in graphs as described in Chapter 4 ________________________________________ GenStat Fifth Edition (Service Pack 1) GenStat Procedure Library Release PL12.1 ________________________________________ 1  %CD 'C:/Documents and Settings/phil.evans/Desktop'  2 "Data taken from File: C:/Documents and Settings/phevans/Desktop/Exp #2.xls" 3 DELETE [Redefine=yes] _stitle_: TEXT _stitle_ 4 READ [print=*;SETNVALUES=yes] _stitle_ 8 PRINT [IPrint=*] _stitle_; Just=Left Data imported from Excel file: C:\Documents and Settings\phevans\Desktop\Exp #2.xls on: 9-Nov-2007 12:50:43 taken from sheet "Check", cells A1:I21 9 10 11 12  DELETE [redefine=yes] Block,Board,Sample,Weathering,Treatment,No,Length,\ Width,Shape FACTOR [modify=yes;nvalues=20;levels=5] Block READ Block; frepresentation=ordinal  Identifier Block 14 15  Missing 0  Levels 2  Values 20  Missing 0  Levels 2  Values 20  Missing 0  Levels 2  FACTOR [modify=yes;nvalues=20;levels=2;labels=!t('Kerf','No-kerf')\ ] Treatment READ Treatment; frepresentation=ordinal  Identifier Treatment 28 29  Values 20  FACTOR [modify=yes;nvalues=20;levels=2;labels=!t('Artificial','Natural')\ ] Weathering READ Weathering; frepresentation=ordinal  Identifier Weathering 24 25 26  Levels 5  FACTOR [modify=yes;nvalues=20;levels=2] Sample READ Sample; frepresentation=ordinal  Identifier Sample 20 21 22  Missing 0  FACTOR [modify=yes;nvalues=20;levels=2] Board READ Board; frepresentation=ordinal  Identifier Board 17 18  Values 20  Values 20  Missing 0  Levels 2  VARIATE [nvalues=20] No READ No  Identifier No  Minimum 67.00  Mean 103.2  Maximum 173.0  Values 20  Missing 0  118  31 32  VARIATE [nvalues=20] Length READ Length  Identifier Length 35 36  Maximum 3114  Values 20  Missing 0  Minimum 4.050  Mean 7.357  Maximum 10.55  Values 20  Missing 0  Maximum 44846  Values 20  Missing 0  VARIATE [nvalues=20] Shape READ Shape  Identifier Shape 43 44  Mean 1973  VARIATE [nvalues=20] Width READ Width  Identifier Width 39 40  Minimum 1087  Minimum 19311  Mean 30220  RESTRICT Block,Board,Sample,Weathering,Treatment,No,Length,Width,Shape  45 "Split-Plot Design." 46 BLOCK Block/Board/Sample 47 TREATMENTS Weathering*Treatment 48 COVARIATE "No Covariate" 49 ANOVA [PRINT=aovtable,information,means; FACT=32; FPROB=yes; PSE=diff,lsd; LSDLEVEL=5]\ 50 No  119  ***** Analysis of variance ***** Variate: No Source of variation  d.f.  s.s.  m.s.  v.r.  Block stratum  4  4927.7  1231.9  0.66  Block.Board stratum Weathering Residual  1 4  72.2 7488.3  72.2 1872.1  0.04 4.82  0.854  Block.Board.Sample stratum Treatment 1 Weathering.Treatment 1 Residual 8  352.8 5.0 3105.2  352.8 5.0 388.2  0.91 0.01  0.368 0.912  Total  19  F pr.  15951.2  ***** Tables of means ***** Variate: No Grand mean  103.2  Weathering Artificial 101.3 Treatment  Kerf 99.0  Natural 105.1  No-kerf 107.4  Weathering Treatment Artificial Natural  Kerf 97.6 100.4  No-kerf 105.0 109.8  *** Standard errors of differences of means *** Table  Weathering  Treatment  Weathering Treatment rep. 10 10 5 s.e.d. 19.35 8.81 21.26 d.f. 4 8 5.71 Except when comparing means with the same level(s) of Weathering 12.46 d.f. 8 *** Least significant differences of means (5% level) *** Table  Weathering  Treatment  Weathering Treatment rep. 10 10 5 l.s.d. 53.72 20.32 52.68 d.f. 4 8 5.71 Except when comparing means with the same level(s) of Weathering 28.73 d.f. 8  120  _______________________________________ GenStat Fifth Edition (Service Pack 1) GenStat Procedure Library Release PL12.1 ________________________________________ 1 2 -3 4 5 9  %CD 'C:/Documents and Settings/phil.evans/Desktop' "Data taken from File: \ C:/Documents and Settings/phevans/Desktop/Copy of Exp #2 detailly.xls" DELETE [Redefine=yes] _stitle_: TEXT _stitle_ READ [print=*;SETNVALUES=yes] _stitle_ PRINT [IPrint=*] _stitle_; Just=Left  Data imported from Excel file: C:\Documents and Settings\phevans\Desktop\Copy of Exp #2 detailly.xls on: 16-Nov-2007 14:26:27 taken from sheet "Length", cells A1:G2065 10 DELETE [redefine=yes] Block,Board,Sample,Weathering,Treatment,Length,Width 11 FACTOR [modify=yes;nvalues=2064;levels=5] Block 12 READ Block; frepresentation=ordinal Identifier Values Missing Levels Block 2064 0 5 68 FACTOR [modify=yes;nvalues=2064;levels=2] Board 69 READ Board; frepresentation=ordinal Identifier Values Missing Levels Board 2064 0 2 125 FACTOR [modify=yes;nvalues=2064;levels=2] Sample 126 READ Sample; frepresentation=ordinal Identifier Sample  Values 2064  Missing 0  Levels 2  182 FACTOR [modify=yes;nvalues=2064;levels=2;labels=!t('Artificial','Natural')\ 183 ] Weathering 184 READ Weathering; frepresentation=ordinal Identifier Weathering 240 241 242  381 382  Missing 0  Levels 2  FACTOR [modify=yes;nvalues=2064;levels=2;labels=!t('Kerf','No-kerf')\ ] Treatment READ Treatment; frepresentation=ordinal  Identifier Treatment 298 299  Values 2064  Values 2064  Missing 0  VARIATE [nvalues=2064] Length READ Length Identifier Minimum Mean Length 8.000 19.12  Levels 2  Maximum 217.0  Values 2064  Missing 0  Skew  VARIATE [nvalues=2064] Width READ Width  Identifier Width  Minimum 0.05000  Mean 0.07129  Maximum 0.9000  Values 2064  Missing 0  Skew  517 RESTRICT Block,Board,Sample,Weathering,Treatment,Length,Width 519 VCOMPONENTS [FIXED=Weathering+Treatment+Weathering.Treatment; FACTORIAL=9] RANDOM=Block+Block.Board+Block.Board.Sample;\ 520 INITIAL=1,1,1; CONSTRAINTS=positive,positive,positive 521 REML [PRINT=model,components,effects,means,waldTests; PSE=differences; MVINCLUDE=*;\ 522 METHOD=AI] Length  121  ***** REML Variance Components Analysis ***** Response Variate : Length Fixed model Random model  : Constant+Weathering+Treatment+Weathering.Treatment : Block+Block.Board+Block.Board.Sample  Number of units  : 2064  * Residual term has been added to model * Sparse algorithm with AI optimisation *** Estimated Variance Components *** Random term  Component  S.e.  7.6 5.0 3.5  8.7 6.0 3.1  Block Block.Board Block.Board.Sample *** Residual variance model *** Term  Factor  Residual  Model(order) Identity  Parameter Sigma2  Estimate  S.e.  276.4  8.6  *** Wald tests for fixed effects *** Fixed term prob  Wald statistic  d.f.  Wald/d.f.  Chi-sq  1 1 1  0.07 0.12 0.25  0.787 0.732 0.616  1  0.25  0.616  * Sequentially adding terms to fixed model Weathering Treatment Weathering.Treatment  0.07 0.12 0.25  * Dropping individual terms from full fixed model Weathering.Treatment  0.25  * Message: chi-square distribution for Wald tests is an asymptotic approximation (i.e. for large samples) and underestimates the probabilities in other cases. *** Table of effects for Constant *** 19.08  Standard error:  1.952  *** Table of effects for Weathering *** Weathering  Artificial 0.00000  Standard error of differences:  Natural 0.08063 2.140  *** Table of effects for Treatment *** Treatment  Kerf 0.0000  No-kerf 0.9566  Standard error of differences:  1.601  122  *** Table of effects for Weathering.Treatment *** Treatment Weathering Artificial Natural  Kerf  No-kerf  0.0000 0.0000  0.0000 -1.1249  Standard error of differences:  2.246  *** Table of predicted means for Constant *** 19.32  Standard error:  1.525  *** Table of predicted means for Weathering *** Weathering  Artificial 19.56  Natural 19.08  Standard error of differences:  1.810  *** Table of predicted means for Treatment *** Treatment  Kerf 19.12  No-kerf 19.51  Standard error of differences:  1.123  *** Table of predicted means for Weathering.Treatment *** Treatment Weathering Artificial Natural  Kerf  No-kerf  19.08 19.16  20.04 18.99  Standard error of differences:  Average Maximum Minimum  Average variance of differences:  1.949 2.140 1.575 3.865  Standard error of differences for same level of factor: Weathering Treatment Average 1.588 2.130 Maximum 1.601 2.140 Minimum 1.575 2.120 Average variance of differences: 2.522 4.537  123  ***** REML Variance Components Analysis ***** Response Variate : Width Fixed model Random model  : Constant+Weathering+Treatment+Weathering.Treatment : Block+Block.Board+Block.Board.Sample  Number of units  : 2064  * Residual term has been added to model * Sparse algorithm with AI optimisation *** Estimated Variance Components *** Random term Component Block Block.Board Block.Board.Sample  0.000091 0.000024 0.000023  S.e. 0.000088 0.000045 0.000036  *** Residual variance model *** Term S.e.  Factor  Residual 0.000150  Model(order)  Parameter  Identity  Sigma2  Estimate 0.00479  *** Wald tests for fixed effects *** Fixed term  Wald statistic  d.f.  Wald/d.f.  Chi-sq prob  * Sequentially adding terms to fixed model Weathering Treatment Weathering.Treatment  5.26 0.61 0.76  1 1 1  5.26 0.61 0.76  0.022 0.435 0.382  1  0.76  0.382  * Dropping individual terms from full fixed model Weathering.Treatment  0.76  * Message: chi-square distribution for Wald tests is an asymptotic approximation (i.e. for large samples) and underestimates the probabilities in other cases. *** Table of effects for Constant *** 0.07822  Standard error:  0.006155  *** Table of effects for Weathering *** Weathering  Artificial 0.000000  Natural -0.014641  Standard error of differences:  0.006236  *** Table of effects for Treatment *** Treatment  Kerf 0.0000000  No-kerf -0.0004133  Standard error of differences:  0.005362  124  *** Table of effects for Weathering.Treatment *** Treatment Weathering Artificial Natural  Kerf  No-kerf  0.000000 0.000000  0.000000 0.006561  Standard error of differences:  0.007511  *** Table of predicted means for Constant *** 0.07234  Standard error:  0.004922  *** Table of predicted means for Weathering *** Weathering  Artificial 0.07802  Natural 0.06666  Standard error of differences:  0.004905  *** Table of predicted means for Treatment *** Treatment  Kerf 0.07090  No-kerf 0.07377  Standard error of differences:  0.003756  *** Table of predicted means for Weathering.Treatment *** Treatment Weathering Artificial Natural  Kerf  No-kerf  0.07822 0.06358  0.07781 0.06973  Standard error of differences:  Average Maximum Minimum  Average variance of differences:  0.005889 0.006236 0.005260 0.00003485  Standard error of differences for same level of factor: Average Maximum Minimum  Weathering 0.005311 0.005362 0.005260  Treatment 0.006177 0.006236 0.006119  125  26 weeks  Analysis of variance Variate: logno Source of variation  d.f.  s.s.  m.s.  v.r.  Board stratum  4  0.071146  0.017786  0.97  Board.Sample stratum Exposure_type Residual  1 4  0.011599 0.073141  0.011599 0.018285  0.63 2.07  0.470  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.000115 0.004694 0.070654  0.000115 0.004694 0.008832  0.01 0.53  0.912 0.487  Total  0.231348  19  F pr.  Tables of means Variate: logno Grand mean 1.808 Exposure_type  Artificial 1.784  Treatment kerfed 1.806  Natural 1.832  unkerfed 1.811  Exposure_type Treatment Artificial Natural  kerfed 1.797 1.814  unkerfed 1.771 1.850  Standard errors of means TableExposure_type TreatmentExposure_type rep. 10 10 e.s.e. 0.0428 0.0297 d.f. 4 8 Except when comparing means with the same level(s) of Exposure_type d.f. Standard errors of differences of means Table Exposure_type  Treatment 5 0.0521 7.88 0.0420 8  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0605 0.0420 0.0736 d.f. 4 8 7.88 Except when comparing means with the same level(s) of Exposure_type 0.0594 d.f. 8 Least significant differences of means (5% level) Table  Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.1679 0.0969 0.1703 d.f. 4 8 7.88 Except when comparing means with the same level(s) of Exposure_type 0.1371 d.f. 8  126  Analysis of variance Variate: logL Source of variation Board stratum  d.f. 4  s.s. 0.041558  m.s. 0.010390  v.r. 0.42  F pr.  1 4  0.008987 0.099057  0.008987 0.024764  0.36 12.79  0.579  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.000091 0.012440 0.015487  0.000091 0.012440 0.001936  0.05 6.43  0.834 0.035  Total  0.177621  Board.Sample stratum Exposure_type Residual  19  Tables of means Variate: logL Grand mean 3.1569 Exposure_type Treatment kerfed 3.1548  Artificial 3.1357  Natural 3.1781  unkerfed 3.1591  Exposure_type Treatment Artificial Natural  kerfed 3.1585 3.1511  unkerfed 3.1129 3.2052  Standard errors of means TableExposure_type TreatmentExposure_type rep. 10 10 e.s.e. 0.04976 0.01391 d.f. 4 8 Except when comparing means with the same level(s) of Exposure_type d.f. Standard errors of differences of means Table Exposure_type  Treatment 5 0.05167 4.64 0.01968 8  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.07038 0.01968 0.07308 d.f. 4 8 4.64 Except when comparing means with the same level(s) of Exposure_type 0.02783 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.19539 0.04537 0.19237 d.f. 4 8 4.64 Except when comparing means with the same level(s) of Exposure_type 0.06417 d.f. 8  127  20 Weeks Analysis of variance Variate: logno Source of variation  d.f.  s.s.  m.s.  v.r.  Board stratum  4  0.097504  0.024376  1.91  Board.Sample stratum Exposure_type Residual  1 4  0.010051 0.050967  0.010051 0.012742  0.79 1.35  0.425  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.000001 0.003445 0.075545  0.000001 0.003445 0.009443  0.00 0.36  0.993 0.563  Total  0.237514  19  F pr.  Tables of means Variate: logno Grand mean 1.762 Exposure_type Treatment kerfed 1.761  Artificial 1.784  Natural 1.739  unkerfed 1.762  Exposure_type Treatment kerfed Artificial 1.797 Natural 1.726 Standard errors of means TableExposure_type TreatmentExposure_type  unkerfed 1.771 1.753  rep. 10 10 e.s.e. 0.0357 0.0307 d.f. 4 8 Except when comparing means with the same level(s) of Exposure_type d.f. Standard errors of differences of means Table Exposure_type  Treatment 5 0.0471 9.51 0.0435 8  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0505 0.0435 0.0666 d.f. 4 8 9.51 Except when comparing means with the same level(s) of Exposure_type 0.0615 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.1402 0.1002 0.1495 d.f. 4 8 9.51 Except when comparing means with the same level(s) of Exposure_type 0.1417 d.f. 8  128  Analysis of variance Variate: logL Source of variation Board stratum  d.f. 4  s.s. 0.031077  m.s. 0.007769  v.r. 0.33  F pr.  1 4  0.002376 0.095459  0.002376 0.023865  0.10 15.01  0.768  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.000481 0.001216 0.012722  0.000481 0.001216 0.001590  0.30 0.76  0.597 0.407  Total  0.143330  Board.Sample stratum Exposure_type Residual  19  Tables of means Variate: logL Grand mean 3.1186 Exposure_type Treatment kerfed 3.1137  Artificial 3.1295  Natural 3.1077  unkerfed 3.1235  Exposure_type Treatment Artificial Natural  kerfed 3.1324 3.0950  unkerfed 3.1266 3.1204  Standard errors of means TableExposure_type TreatmentExposure_type rep. 10 10 e.s.e. 0.04885 0.01261 d.f. 4 8 Except when comparing means with the same level(s) of Exposure_type d.f. Standard errors of differences of means Table Exposure_type  Treatment 5 0.05045 4.54 0.01783 8  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.06909 0.01783 0.07135 d.f. 4 8 4.54 Except when comparing means with the same level(s) of Exposure_type 0.02522 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.19181 0.04113 0.18913 d.f. 4 8 4.54 Except when comparing means with the same level(s) of Exposure_type 0.05816 d.f. 8  129  16 Weeks Analysis of variance Variate: logno Source of variation Board stratum Board.Sample stratum Exposure_type Residual  d.f. 4  s.s. 0.120447  m.s. 0.030112  v.r. 2.67  F pr.  1 4  0.089586 0.045183  0.089586 0.011296  7.93 1.55  0.048  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.006581 0.000544 0.058459  0.006581 0.000544 0.007307  0.90 0.07  0.370 0.792  Total  0.320801  19  Tables of means Variate: logno Grand mean 1.717 Exposure_type Treatment kerfed 1.735  Artificial 1.784  Natural 1.650  unkerfed 1.699  Exposure_type Treatment Artificial Natural  kerfed 1.797 1.674  Standard errors of means Table Exposure_type  unkerfed 1.771 1.627  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0336 0.0270 0.0431 d.f. 4 8 8.97 Except when comparing means with the same level(s) of Exposure_type 0.0382 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0475 0.0382 0.0610 d.f. 4 8 8.97 Except when comparing means with the same level(s) of Exposure_type 0.0541 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.1320 0.0882 0.1381 d.f. 4 8 8.97 Except when comparing means with the same level(s) of Exposure_type 0.1247 d.f. 8  130  Analysis of variance Variate: logL Source of variation Board stratum  d.f. 4  s.s. 0.051679  m.s. 0.012920  v.r. 0.67  F pr.  1 4  0.065129 0.076568  0.065129 0.019142  3.40 13.51  0.139  1 1 8  0.010168 0.000001 0.011339  0.010168 0.000001 0.001417  7.17 0.00  0.028 0.976  19  0.214884  Board.Sample stratum Exposure_type Residual Board.Sample.Specimen stratum Treatment Exposure_type.Treatment Residual Total  Message: the following units have large residuals. Board 1 Sample 2 Specimen 1 -0.0522 s.e. 0.0238 Board 1 Sample 2 Specimen 2 0.0522 s.e. 0.0238 Tables of means Variate: logL Grand mean 3.0787 Exposure_type Treatment kerfed 3.1012  Artificial 3.1357  Natural 3.0216  unkerfed 3.0561  Exposure_type Treatment Artificial Natural  kerfed 3.1585 3.0439  Standard errors of means Table Exposure_type  unkerfed 3.1129 2.9993  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.04375 0.01191 0.04534 d.f. 4 8 4.60 Except when comparing means with the same level(s) of Exposure_type 0.01684 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.06187 0.01684 0.06412 d.f. 4 8 4.60 Except when comparing means with the same level(s) of Exposure_type 0.02381 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.17178 0.03883 0.16922 d.f. 4 8 4.60 Except when comparing means with the same level(s) of Exposure_type 0.05491 d.f. 8  131  12 Weeks  Analysis of variance Variate: logno Source of variation Board stratum  d.f. 4  s.s. 0.122120  m.s. 0.030530  v.r. 1.99  F pr.  1 4  0.414290 0.061240  0.414290 0.015310  27.06 1.77  0.007  1 1 8  0.004285 0.000059 0.069291  0.004285 0.000059 0.008661  0.49 0.01  0.502 0.936  19  0.671284  Board.Sample stratum Exposure_type Residual Board.Sample.Specimen stratum Treatment Exposure_type.Treatment Residual Total Tables of means Variate: logno Grand mean 1.640 Exposure_type Treatment kerfed 1.655  Artificial 1.784  Natural 1.496  unkerfed 1.625  Exposure_type Treatment Artificial Natural  kerfed 1.797 1.513  Standard errors of means Table Exposure_type  unkerfed 1.771 1.480  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0391 0.0294 0.0490 d.f. 4 8 8.45 Except when comparing means with the same level(s) of Exposure_type 0.0416 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0553 0.0416 0.0692 d.f. 4 8 8.45 Except when comparing means with the same level(s) of Exposure_type 0.0589 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.1536 0.0960 0.1582 d.f. 4 8 8.45 Except when comparing means with the same level(s) of Exposure_type 0.1357 d.f. 8  132  Analysis of variance Variate: logL Source of variation  d.f.  s.s.  m.s.  v.r.  Board stratum  4  0.045515  0.011379  0.48  Board.Sample stratum Exposure_type Residual  1 4  0.475619 0.095202  0.475619 0.023800  19.98 5.12  0.011  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.008668 0.000079 0.037181  0.008668 0.000079 0.004648  1.87 0.02  0.209 0.899  Total  0.662263  19  F pr.  Tables of means Variate: logL Grand mean 2.982 Exposure_type Treatment kerfed 3.002  Artificial 3.136  Natural 2.827  unkerfed 2.961  Exposure_type Treatment Artificial Natural  kerfed 3.159 2.846  Standard errors of means Table Exposure_type  unkerfed 3.113 2.808  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0488 0.0216 0.0533 d.f. 4 8 5.61 Except when comparing means with the same level(s) of Exposure_type 0.0305 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0690 0.0305 0.0754 d.f. 4 8 5.61 Except when comparing means with the same level(s) of Exposure_type 0.0431 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.1915 0.0703 0.1877 d.f. 4 8 5.61 Except when comparing means with the same level(s) of Exposure_type 0.0994 d.f. 8  133  8 Weeks Analysis of variance Variate: logno Source of variation d.f. Board stratum 4 Board.Sample stratum Exposure_type 1 Residual 4 Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8 Total  19  s.s. 0.053870  m.s. 0.013467  v.r. 0.35  F pr.  1.434481 0.155456  1.434481 0.038864  36.91 5.93  0.004  0.004197 0.000049 0.052458  0.004197 0.000049 0.006557  0.64 0.01  0.447 0.933  1.700511  Message: the following units have large residuals. Board 3 Sample 2 Specimen 1 0.104 s.e. 0.051 Board 3 Sample 2 Specimen 2 Tables of means Variate: logno Grand mean 1.516  -0.104  s.e. 0.051  Exposure_type  Artificial Natural 1.784 1.248 Treatment kerfed unkerfed 1.531 1.502 Exposure_type Treatment kerfed unkerfed Artificial 1.797 1.771 Natural 1.264 1.232 Standard errors of means Table Exposure_type  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0623 0.0256 0.0674 d.f. 4 8 5.39 Except when comparing means with the same level(s) of Exposure_type 0.0362 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0882 0.0362 0.0953 d.f. 4 8 5.39 Except when comparing means with the same level(s) of Exposure_type 0.0512 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2448 0.0835 0.2398 d.f. 4 8 5.39 Except when comparing means with the same level(s) of Exposure_type 0.1181 d.f. 8  134  Analysis of variance Variate: loglength Source of variation  d.f.  s.s.  m.s.  v.r.  Board stratum  4  0.084409  0.021102  0.41  Board.Sample stratum Exposure_type Residual  1 4  1.122307 0.207942  1.122307 0.051985  21.59 20.10  0.010  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.007551 0.000228 0.020688  0.007551 0.000228 0.002586  2.92 0.09  0.126 0.774  Total  1.443125  19  F pr.  Tables of means Variate: loglength Grand mean 2.899 Exposure_type  Artificial 3.136  Treatment kerfed 2.918  unkerfed 2.879  Exposure_type Treatment Artificial Natural Standard errors of means Table  Natural 2.662  kerfed 3.159 2.678  Exposure_type  unkerfed 3.113 2.646  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0721 0.0161 0.0739 d.f. 4 8 4.40 Except when comparing means with the same level(s) of Exposure_type 0.0227 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.1020 0.0227 0.1045 d.f. 4 8 4.40 Except when comparing means with the same level(s) of Exposure_type 0.0322 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2831 0.0524 0.2799 d.f. 4 8 4.40 Except when comparing means with the same level(s) of Exposure_type 0.0742 d.f. 8  135  4 Weeks Analysis of variance Variate: logno Source of variation Board stratum Board.Sample stratum Exposure_type Residual  d.f. 4  s.s. 0.096140  m.s. 0.024035  v.r. 0.45  F pr.  1 4  2.682675 0.214271  2.682675 0.053568  50.08 9.01  0.002  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.001799 0.010044 0.047580  0.001799 0.010044 0.005947  0.30 1.69  0.597 0.230  Total  3.052508  19  Tables of means Variate: logno Grand mean 1.418 Exposure_type Treatment kerfed 1.408  Artificial 1.784  Natural 1.052  unkerfed 1.427  Exposure_type Treatment Artificial Natural  kerfed 1.797 1.020  Standard errors of means Table Exposure_type  unkerfed 1.771 1.083  Exposure_type Treatment rep. 10 10 5 e.s.e. 0.0732 0.0244 0.0771 d.f. 4 8 4.91 Except when comparing means with the same level(s) of Exposure_type 0.0345 d.f. 8 Standard errors of differences of means Table Exposure_type  Treatment  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.1035 0.0345 0.1091 d.f. 4 8 4.91 Except when comparing means with the same level(s) of Exposure_type 0.0488 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2874 0.0795 0.2821 d.f. 4 8 4.91 Except when comparing means with the same level(s) of Exposure_type 0.1125 d.f. 8  136  Analysis of variance Variate: loglength Source of variation Board stratum  d.f. 4  s.s. 0.137275  m.s. 0.034319  v.r. 0.64  F pr.  1 4  2.222649 0.213707  2.222649 0.053427  41.60 23.25  0.003  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.000316 0.014343 0.018381  0.000316 0.014343 0.002298  0.14 6.24  0.721 0.037  Total  2.606672  Board.Sample stratum Exposure_type Residual  19  Tables of means Variate: loglength Grand mean 2.802 Exposure_type Treatment kerfed 2.798  Artificial 3.136  Natural 2.469  unkerfed 2.806  Exposure_type Treatment Artificial Natural  kerfed 3.159 2.438  unkerfed 3.113 2.500  Standard errors of means TableExposure_type TreatmentExposure_type rep. 10 10 e.s.e. 0.0731 0.0152 d.f. 4 8 Except when comparing means with the same level(s) of Exposure_type d.f. Standard errors of differences of means Table Exposure_type  Treatment 5 0.0746 4.35 0.0214 8  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.1034 0.0214 0.1056 d.f. 4 8 4.35 Except when comparing means with the same level(s) of Exposure_type 0.0303 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2870 0.0494 0.2841 d.f. 4 8 4.35 Except when comparing means with the same level(s) of Exposure_type 0.0699 d.f. 8  137  2 Weeks  Analysis of variance Variate: logno Source of variation d.f. Board stratum 4 Board.Sample stratum Exposure_type 1 Residual 4 Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8 Total  19  s.s. 0.06070  m.s. 0.01518  v.r. 0.39  F pr.  5.49287 0.15625  5.49287 0.03906  140.62 3.35  <.001  0.01573 0.03357 0.09335  0.01573 0.03357 0.01167  1.35 2.88  0.279 0.128  5.85248  Tables of means Variate: logno Grand mean 1.260 Exposure_type Treatment kerfed 1.232  Artificial 1.784  Natural 0.736  unkerfed 1.288  Exposure_type Treatment Artificial Natural  kerfed 1.797 0.667  Standard errors of differences of means Table Exposure_type  unkerfed 1.771 0.805  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.0884 0.0483 0.1007 d.f. 4 8 6.46 Except when comparing means with the same level(s) of Exposure_type 0.0683 d.f. 8 Least significant differences of means (5% level) Table Exposure_type Treatment Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2454 0.1114 0.2423 d.f. 4 8 6.46 Except when comparing means with the same level(s) of Exposure_type 0.1575 d.f. 8  138  Analysis of variance Variate: loglength Source of variation  d.f.  s.s.  m.s.  v.r.  Board stratum  4  0.131968  0.032992  0.65  Board.Sample stratum Exposure_type Residual  1 4  4.446854 0.203065  4.446854 0.050766  87.59 7.31  <.001  Board.Sample.Specimen stratum Treatment 1 Exposure_type.Treatment 1 Residual 8  0.002109 0.021882 0.055547  0.002109 0.021882 0.006943  0.30 3.15  0.597 0.114  Total  4.861425  19  F pr.  Tables of means Variate: loglength Grand mean 2.664 Exposure_type  Artificial 3.136  Treatment kerfed 2.654  Natural 2.193  unkerfed 2.674  Exposure_type Treatment Artificial Natural  kerfed 3.159 2.149  unkerfed 3.113 2.236  Standard errors of differences of means Table  Exposure_type  Treatment  Exposure_type Treatment rep. 10 10 5 s.e.d. 0.1008 0.0373 0.1074 d.f. 4 8 5.12 Except when comparing means with the same level(s) of Exposure_type 0.0527 d.f. 8 Least significant differences of means (5% level) Table  Exposure_type  Treatment  Exposure_type Treatment rep. 10 10 5 l.s.d. 0.2798 0.0859 0.2742 d.f. 4 8 5.12 Except when comparing means with the same level(s) of Exposure_type 0.1215 d.f. 8  139  Data from 2nd Experiment (Chapter 4) (for Appendix 2) Block  Board  Sample  Weathering  Total Check Length (mm)  Treatment No  L/W  Width (mm)  ratio  1  1  1  Natural  Kerf  73  1087  4.05  19,635  1  1  2  Natural  No-kerf  111  1535  6  28,965  1  2  1  Artificial  Kerf  151  2503  8.65  44,846  1  2  2  Artificial  No-kerf  173  2360  10.55  40,691  2  1  1  Natural  Kerf  82  1525  5.3  24,095  2  1  2  Natural  No-kerf  104  2230  7.6  33,467  2  2  1  Artificial  Kerf  67  1320  5.55  19,553  2  2  2  Artificial  No-kerf  68  1494  6.5  19,311  3  1  1  Natural  Kerf  104  2419  7.35  33,790  3  1  2  Natural  No-kerf  97  1889  7  28,372  3  2  1  Artificial  Kerf  102  1387  7.7  21,261  3  2  2  Artificial  No-kerf  72  1445  5.35  21,692  4  1  1  Natural  Kerf  110  2033  7  32,222  4  1  2  Natural  No-kerf  142  2686  9.85  40,766  4  2  1  Artificial  Kerf  77  1236  4.85  21,107  4  2  2  Artificial  No-kerf  91  1802  6.9  26,404  5  1  1  Natural  Kerf  133  2725  8.65  43,775  5  1  2  Natural  No-kerf  95  2019  7.6  28,907  5  2  1  Artificial  Kerf  91  2650  10.3  32,326  5  2  2  Artificial  No-kerf  121  3114  10.4  43,225  140  Distortion (mm)  Block  Board  Sample  Weathering  Treatment  1  1  1  Natural  Kerf  0.17  1.84  0.31  1  1  2  Natural  No-kerf  0.14  1.26  0.04  1  2  1  Artificial  Kerf  0.57  1  2  2  Artificial  No-kerf  0.7  0.32  0.27  2  1  1  Natural  Kerf  -0.16  0.80  0.28  2  1  2  Natural  No-kerf  1.42  1.78  0.09  2  2  1  Artificial  Kerf  0.49  0.30  0.06  2  2  2  Artificial  No-kerf  0.36  0.58  0.11  3  1  1  Natural  Kerf  -1.21  0.75  -0.14  3  1  2  Natural  No-kerf  0.98  2.18  0.7  3  2  1  Artificial  Kerf  1.57  0.98  0.12  3  2  2  Artificial  No-kerf  1.12  0.90  -0.07  4  1  1  Natural  Kerf  -0.24  1.06  0.11  4  1  2  Natural  No-kerf  -1.07  0.60  -0.15  4  2  1  Artificial  Kerf  0.53  1.88  0.14  4  2  2  Artificial  No-kerf  1.2  (0.39)  0.91  5  1  1  Natural  Kerf  0.65  0.10  0.15  5  1  2  Natural  No-kerf  0.45  1.24  -0.05  5  2  1  Artificial  Kerf  0.74  0.76  0.1  5  2  2  Artificial  No-kerf  1.68  0.15  0.05  Cup C  Twist  -0.16  Bow  0.11  141  Exposure  Sample  Specimen  Replicate  1  1  1  1  Natural  Kerfed  1  3  73.35  1  1  1  2  Natural  Kerfed  2  6  141.16  1  1  1  3  Natural  Kerfed  3  12  240.81  1  1  1  4  Natural  Kerfed  4  36  635.24  1  1  1  5  Natural  Kerfed  5  48  871.90  1  1  1  6  Natural  Kerfed  6  52  982.62  1  1  1  7  Natural  Kerfed  7  55  1,036.59  1  1  2  1  Natural  Unkerfed  1  6  110.72  1  1  2  2  Natural  Unkerfed  2  9  184.07  1  1  2  3  Natural  Unkerfed  3  14  269.87  1  1  2  4  Natural  Unkerfed  4  26  395.81  1  1  2  5  Natural  Unkerfed  5  47  804.08  1  1  2  6  Natural  Unkerfed  6  61  1,054.58  1  1  2  7  Natural  Unkerfed  7  73  1,209.59  1  2  1  1  Artificial  Kerfed  1  31  751.49  1  2  1  2  Artificial  Kerfed  2  44  957.70  1  2  1  3  Artificial  Kerfed  3  62  1,274.63  1  2  1  4  Artificial  Kerfed  4  71  1,554.19  1  2  1  5  Artificial  Kerfed  5  86  1,765.94  1  2  2  1  Artificial  Unkerfed  1  40  888.51  1  2  2  2  Artificial  Unkerfed  2  62  1,179.14  1  2  2  3  Artificial  Unkerfed  3  76  1,302.31  1  2  2  4  Artificial  Unkerfed  4  92  1,876.66  1  2  2  5  Artificial  Unkerfed  5  106  2,021.97  2  1  1  1  Natural  Kerfed  1  5  196.52  2  1  1  2  Natural  Kerfed  2  10  315.54  2  1  1  3  Natural  Kerfed  3  22  712.74  2  1  1  4  Natural  Kerfed  4  31  761.18  2  1  1  5  Natural  Kerfed  5  47  1,080.88  2  1  1  6  Natural  Kerfed  6  54  1,162.53  2  1  1  7  Natural  Kerfed  7  61  1,435.17  2  1  2  1  Natural  Unkerfed  1  7  221.43  2  1  2  2  Natural  Unkerfed  2  12  406.89  2  1  2  3  Natural  Unkerfed  3  15  505.15  2  1  2  4  Natural  Unkerfed  4  21  602.03  2  1  2  5  Natural  Unkerfed  5  32  1,100.25  2  1  2  6  Natural  Unkerfed  6  45  1,432.41  2  1  2  7  Natural  Unkerfed  7  65  1,815.00  2  2  1  1  Artificial  Kerfed  1  25  568.81  2  2  1  2  Artificial  Kerfed  2  32  721.05  2  2  1  3  Artificial  Kerfed  3  44  1,096.10  2  2  1  4  Artificial  Kerfed  4  46  1,169.45  2  2  1  5  Artificial  Kerfed  5  48  1,227.58  type  Treatment  Time  Total Check  Board  Number  Length (mm)  142  Board  Sample  Specimen  Replicate  2  2  2  1  2  2  2  2  2  2  2  2  2 3  Exposure  Treatment  Time  Artificial  Unkerfed  2  Artificial  3  Artificial  2  4  2  2  1  1  3  1  3 3  type  Total Check Number  Length (mm)  1  18  513.45  Unkerfed  2  23  632.47  Unkerfed  3  31  718.28  Artificial  Unkerfed  4  36  849.76  5  Artificial  Unkerfed  5  39  1,071.19  1  Natural  Kerfed  1  6  163.31  1  2  Natural  Kerfed  2  11  269.87  1  1  3  Natural  Kerfed  3  18  470.55  1  1  4  Natural  Kerfed  4  24  604.79  3  1  1  5  Natural  Kerfed  5  38  1,157.00  3  1  1  6  Natural  Kerfed  6  43  1,313.38  3  1  1  7  Natural  Kerfed  7  63  1,346.60  3  1  2  1  Natural  Unkerfed  1  6  131.48  3  1  2  2  Natural  Unkerfed  2  11  239.43  3  1  2  3  Natural  Unkerfed  3  18  412.42  3  1  2  4  Natural  Unkerfed  4  32  755.65  3  1  2  5  Natural  Unkerfed  5  42  970.16  3  1  2  6  Natural  Unkerfed  6  52  1,364.59  3  1  2  7  Natural  Unkerfed  7  62  1,577.72  3  2  1  1  Artificial  Kerfed  1  36  646.31  3  2  1  2  Artificial  Kerfed  2  44  795.78  3  2  1  3  Artificial  Kerfed  3  57  995.07  3  2  1  4  Artificial  Kerfed  4  61  1,049.05  3  2  1  5  Artificial  Kerfed  5  72  1,186.06  3  2  2  1  Artificial  Unkerfed  1  18  617.25  3  2  2  2  Artificial  Unkerfed  2  22  715.51  3  2  2  3  Artificial  Unkerfed  3  31  1,101.64  3  2  2  4  Artificial  Unkerfed  4  36  1,192.98  3  2  2  5  Artificial  Unkerfed  5  42  1,054.58  4  1  1  1  Natural  Kerfed  1  3  94.11  4  1  1  2  Natural  Kerfed  2  10  279.56  4  1  1  3  Natural  Kerfed  3  17  466.40  4  1  1  4  Natural  Kerfed  4  31  694.75  4  1  1  5  Natural  Kerfed  5  40  1,082.26  4  1  1  6  Natural  Kerfed  6  49  1,346.60  4  1  1  7  Natural  Kerfed  7  65  1,493.30  4  1  2  1  Natural  Unkerfed  1  6  185.45  4  1  2  2  Natural  Unkerfed  2  11  333.54  4  1  2  3  Natural  Unkerfed  3  16  471.93  4  1  2  4  Natural  Unkerfed  4  35  748.73  4  1  2  5  Natural  Unkerfed  5  44  953.55  4  1  2  6  Natural  Unkerfed  6  78  1,453.16  4  1  2  7  Natural  Unkerfed  7  99  17,778.40  4  2  1  1  Artificial  Kerfed  1  20  377.82  4  2  1  2  Artificial  Kerfed  2  23  708.59  143  Exposure  Board  Sample  Specimen  Replicate  4  2  1  3  Artificial  Kerfed  4  2  1  4  Artificial  4  2  1  5  Artificial  4  2  2  1  4  2  2  4  2  2  4  2  4 5  type  Treatment  Time  Total Check Number  Length (mm)  3  41  970.16  Kerfed  4  47  1,234.50  Kerfed  5  57  1,368.74  Artificial  Unkerfed  1  16  344.61  2  Artificial  Unkerfed  2  21  487.16  3  Artificial  Unkerfed  3  28  614.48  2  4  Artificial  Unkerfed  4  39  815.16  2  2  5  Artificial  Unkerfed  5  59  1,013.06  1  1  1  Natural  Kerfed  1  8  251.88  5  1  1  2  Natural  Kerfed  2  19  462.24  5  1  1  3  Natural  Kerfed  3  26  651.85  5  1  1  4  Natural  Kerfed  4  44  837.30  5  1  1  5  Natural  Kerfed  5  68  1,404.73  5  1  1  6  Natural  Kerfed  6  72  1,478.08  5  1  1  7  Natural  Kerfed  7  86  1,902.95  5  1  2  1  Natural  Unkerfed  1  7  253.27  5  1  2  2  Natural  Unkerfed  2  20  527.29  5  1  2  3  Natural  Unkerfed  3  24  639.39  5  1  2  4  Natural  Unkerfed  4  41  817.92  5  1  2  5  Natural  Unkerfed  5  49  1,212.35  5  1  2  6  Natural  Unkerfed  6  52  1,335.53  5  1  2  7  Natural  Unkerfed  7  61  1,723.04  5  2  1  1  Artificial  Kerfed  1  15  484.39  5  2  1  2  Artificial  Kerfed  2  22  757.03  5  2  1  3  Artificial  Kerfed  3  36  1,155.61  5  2  1  4  Artificial  Kerfed  4  43  1,443.48  5  2  1  5  Artificial  Kerfed  5  57  1,763.17  5  2  2  1  Artificial  Unkerfed  1  19  571.58  5  2  2  2  Artificial  Unkerfed  2  28  740.42  5  2  2  3  Artificial  Unkerfed  3  39  1,161.15  5  2  2  4  Artificial  Unkerfed  4  58  1,440.71  5  2  2  5  Artificial  Unkerfed  5  70  1,586.03  144  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0067135/manifest

Comment

Related Items