SAWING TO REDUCE WARP OF LODGEPOLE PINE STUDS

Transcription

1 U.S.D.A. FOREST SERVICE RESEARCH PAPER FPL 102 MARCH 1969 U.S. Department of Agriculture/Forest Service/Forest Products Laboratory/Madison, Wis. SAWING TO REDUCE WARP OF LODGEPOLE PINE STUDS

2 SUMMARY Six hundred and eighthy lodgepole pine logs with 6- through 12-inch diameters were sawned into 2- by 4-inch studs to investigate the relation of five variables to warp. Included in the variables were sawing method, position of the log in the tree, position of the stud in the log, rotational position of the log's cross sectional eccentricity, and presence of visible compression wood. The FPL improved Scragg method proved slightly superior to the standard Scragg method, and both were substantially better than the conventional method in yield of quality studs. Studs from butt logs showed more bow and crook, but less twist than those from upper logs. When all forms of warp are considered, there is little difference between yields from butt and upper logs. Bow and crook were more severe in studs from the area near the center of log, the inner area, than in those from the area from the bark, the outer area. Twist appeared unrelated to position of the stud in the log. There appeared to be a relationship between the placement of the sawing pattern and the crosssectional eccentricity. The vertical positioning produced more upper grade studs. The presence of visibly evident compression wood on one or both ends of the log showed only a very minor relationship to warp and raised the question of the validity of identifying compression of wood by visualdetection mothods only. From the results of the investigation, reccomendations are made to producers of lodgepole pine studs. ACKNOWLEDGMENT The author wishes to acknowledge personnel of the Intermoutian Forest and Range Experiment Station, U.S. Forest Service, Ogden, Utah, and the Elk Studs Company, West Yellowstone, Montana, for their cooperation in procuring and preparing the logs for this study. Acknowledge is also made to Statistician Frank Freese, the Forest Products Laboratory, for providing assistance in the design of the study and in the statistical analysis of the results.

3 SAWING TO REDUCE WARP OF LODGEPOLE PINE STUDS By HIRAM HALLOCK, Technologist Forest Products Laboratory 1 Forest Service U.S Department of Agriculture lntroduction Each year the softwood lumber industry In the past, researchers have tried to reduce becomes increasingly dependent on second-growth warp through improved lumber-piling methods, timber and on the old-growth species of inherently inclusive of edge stacking of dimension lumber small diameter. Use of lumber from these and drying under restraint. Air-drying practices sources has resulted in an increase of warp in and kiln schedules have been thoroughly investilumber. This is especially true of studs (nomi- gated. Although the problem has been helped by nally 2 inches by 4 inches by 8 feet). these studies, it continues to be serious. In recent years several softwood lumber- For these reasons, the significance of sawing grading bureaus have found it advisable to issue to prevent warp is being investigated at the special grading rules for studs in which emphasis Forest Products Laboratory (FPL). This report is on straightness or on freedom from warp. The is concerned with lodgepole pine (Pinus contorta application of these rules has sharply reduced Dougl.) from which a substantial volume of studs the amount of warp allowable for studs. Although is produced annually. In an earlier study, examinathis development has benefited the ultimate con- tion was made of the effect of sawing methods sumer, it has resulted in a much higher percent- and other factors on the warp of studs from age of the lower grades being recovered at the loblolly pine (Pinus taeda L.) (5). 2 sawmill. 1 Maintained at Madison, Wis., in cooperation with the University of Wisconsin. 2 Underlined numbers in parentheses refer to Literature Cited at the end of this report.

4 Forms of Warp Warp (13) denotes any deviation of a piece of lumber from a true or plane surface; bow, crook, twist, and cup are forms of warp. Bow is a distortion in a board from lengthwise flatness but not across its faces; crook, the deviation of a piece of lumber edgewise from a straight line from end to end; twist, the turning or winding of the edges in a manner that the four corners of any face are not in the same plane; and cup, a deviation flatwise from a straight line across the width of a board. Bow, crook, and twist occur frequently and to a varying degree in nominal 2- by 4-inch studs. Cup is seldom present in sufficient degree to affect either the grade or the use of the stud. Review of Literature Many investigators of the relation of lumber characteristics to warp or degrade have recognized that the sawing method should be important in reducing potential warp. Cockrell (1) states In butt logs, the combination of high longitudinal shrinkage near the pith and slight shrinkage or elongation of the wood further out can cause extreme counter tension in seasoning which would result in pronounced warping of boards cut with one edge along the pith. Although the cross grain surrounding knots would, to a lesser degree, affect the slight shrinkage or elongation of the outer wood, no lumber can be safely cut with the pith along one edge without increasing the risk of degrade in seasoning. It would, therefore, be good practice to saw logs, and especially butt logs, so that the pith is approximately in the center of the board or timber. Differential longitudinal shrinkage during drying has long been recognized as the primary cause of bow and crook, and frequently is a contributing cause of twist. Koehler and others at the Forest Products Laboratory (14) have shown that the normal longitudinal shrinkage of wood is from 0.1 to 0.3 percent. They have noted abnormal longitudinal shrinkage up to 5.78 percent in ponderosa pine (Pinus ponderosa Laws.) compression wood; this is equivalent to a shortening of more than 5 inches in an 8-foot stud. Excessive longitudinal shrinkage in areas of non-normal wood have also been reported by numerous other writers including Cockrell (1, 2), Du Toit (3), King (6, 7), Paul (10), Paul and Sweet (11), and Pillow (12). King (6, 7) concluded that the application of water-repellent chemicals helped reduce distortion in stored southern pine studs but estimated that 80 percent of the warping problem could be eliminated by proper selection of logs and by changing the patterns of sawing. Kloot and Page (9) in working with radiata pine (Pinus radiata D. Don) suggested that sawing patterns that eliminate the central portion of the log should reduce the warping problem. Paul and Sweet (11) made three recommendations for sawing patterns in southern yellow pine: (1) Control sawing pattern to reduce mixing fast and slow growth or compression and normal wood, and possibly employ taper sawing: (2) cut center of log into boards if wide ringed; and (3) saw crooked logs to have the wide lumber face at right angles to the plane of the crook. Zobel (15) quoted correspondence from Jennings of Australia as follows: The existence of the core (juvenile wood formed the first years of growth) must be recognized in the sawing pattern, because it is fundamentally unstable in seasoning, low in mechanical strength and, consequently, very low in value. Zobel also stated, Sawing of lumber with core wood included invites trouble since it dries differently from other wood, having excessive longitudinal shrinkage and other difficulties. Kotok (8) investigated the relationship of seven tree characteristics-- general sweep, spiral grain, short butt crook, eccentric pith, stern crook, fluted butt and forked stem--on the grade yields of studs sawed from several diameter classes. He found the presence of eccentric pith, stem crook, fluted butt, or forkedness positively correlated with reduced yields of upper-grade studs. He also reported a positive correlation between increasing log diameter and increasing yield of upper grades. A thorough search of the literature indicates that the first study in which control of the internal characteristics of sawed lumber was attempted by FPL 102 2

5 controlling and altering the sawing pattern was conducted by Hallock (5). In1965 Hallock reported a significant relationship between sawing method and subsequent warp in studs. However, sawing methods that reduced one aspect of warp frequently increased other aspects. Because the overriding cause of warp degrade was crook for the species studied, loblolly pine, the sawing method that reduced crook the most was determined the best overall method. Conduct of Study Scope The subject of this report is the relationship of warp in lodgepole pine studs to sawing methods, to log cross-sectional eccentricity, to position of the stud in the log, to log diameter, to log position in the tree, and to presence of compression wood. An additional subject is the effect on warp of long-term storage of strapped packages of lodgepole pine studs. Study Design This study was designed to have 5 percent differences in yield of grade 1 studs between any two treatments statistically significant at the 0.05 level. To keep the size of the investigation within manageable levels for number of logs, the investigation was subdivided into three substudies. Substudy 1 was of 4 by 2 by 2 by 2 factorial design (four diameters by two log positions by two eccentricities by two saw methods). Substudy 2 was designed to compare a third sawing method with the two methods investigated in substudy 1 and was of 3 by 2 by 2 by 1 factorial design (three sawing methods by two log positions by two eccentricities by one diameter). Substudy 3 was designed to investigate the effect of the presence of compression wood on warp. It was of 2 by 2 by 2 by 2 by 1 factorial design (two classes of compression wood by two diameter classes by two sawing methods by two eccentricities by one log position). In each of the substudies, 15 logs for each factorial combination were selected; these were further randomly divided into three replications of which each contained five logs. Variables A primary objective was to provide information that could be used by sawmill personnel to produce a product with a lower potential for warp. Thus the selection of variables was confined to the characteristics that are readily identifiable visually. Log diameter.--most lodgepole pine sawmills that produce studs are using small logs in the diameter range 6 through 12 inches measured at the top d.i.b. (diameter inside bark). For the study as a whole this diameter range was grouped into four classes: 6 inches (5.1 to 7.0), 8 inches (7.1 to 9.0), 10 inches (9.1 to 11.0), and 12 inches (11.0 to 13.0). In substudy 1, all diameter classes were used; in substudy 2, only the 12-inch class; and in substudy 3, only the 8- and 10-inch-diameter classes. Sawing method. --The earlier study on loblolly pine (5) included four sawing methods (fig. 1). Because Method II is not used commercially and does not yield as high grades as the present method used by industry (Method III), it was decided to drop Method II from this study. Most mills that produce studs in quantity use one of two basic sawing methods--conventional or Scragg. The minimum conventional mill combines a standard log carriage with either a circular or a band headrig, and usually has secondary breakdown equipment such as an edger and a resaw or a gang. The log is placed on the carriage and, by a certain sequence of sawing and turning, it is reduced either to the final product or to flitches and cants for further remanufacture by whatever secondary equipment is available. This basic system is suitable for producing studs by anyone of the three methods to be described. The Scragg system is suitable for producing studs only by Methods III and IV (fig. 1). With the Scragg system, the log passes between successive pairs of parallel (usually circular) saws that remove slabs, 2-inch flitches, and leave a 4-inch cant from the approximate geometric center of the log. The 2-inch flitches are then passed through a circular gang and ripped into 3

6 Figure 1.--Four sawing methods on 12-inch diameter class that can be used to reduce logs to studs. In Methods I and III taper and excess wood (shaded area) are in the slab, whereas in Methods II and IV they are mainly concentrated in a wedge from the pith area. stock 4 inches wide. The 4-inch cant is also passed through another section of the circular gang and ripped into stock 2 inches thick. Thus the log is usually sawed in a more or less continuous process to yield only studs. This system is ideal for Method III and with minor alterations, for Method IV. Method I (fig. 1) is commonly used at conventional mills with carriage and band or circular headrigs. By this method any log large enough to develop two adjacent 4-inch cants can be sawed. In substudy 2 this method was used to saw one-third of the 12-inch d.i.b. logs. Characteristically it produces two parallel and adjacent 4-inch cants that tend to have the pith centrally located on the inner wide faces. Each cant is then ripped parallel to its longitudinal axis into 2- by 4-inch stock. The taper of the log and all wood not of sufficient size to yield an additional stud is removed in the slabs and edgings. With this method all of the juvenile core (less sawdust) becomes a part of the studs. Method III (Scragg method) is suitable for logs with diameters of more than 5 inches. This method is used to produce studs at all Scragg mills and occasionally at conventional mills. When applied to logs under approximately 9.5 inches, a single 4-inch cant is produced that tends to have the pith boxed in its approximate geometric center. With most logs of 10- and 12-inch diameter classes, an additional 2-inch flitch is developed parallel and adjacent to each wide face of the central 4-inch cant. In this method, the 4-inch cant and the 2-inch flitches are sawed in the manner described for Scragg mills; all sawing is approximately parallel to the central longitudinal axis of the cant or flitch. Thus log taper and excess wood in the cant or flitch are removed as tapered slabs and edgings. The entire juvenile core less sawdust is included in the lumber. This method has an advantage, theoretically at least, over Method I; it tends to confine the core to fewer studs and to yield a higher percentage of studs with balanced growth stresses in the 4-inch plane. Method IV (FPL improved Scragg method) is a modification of Method III that, to our knowledge, has never been used commercially. The 2-inch flitches and 4-inch cant are produced exactly as in Method III. If the 2-inch flitch is FPL 102 4

7 wide enough to yield two studs, these studs are ripped parallel to the adjacent bark edges, and excess wood and taper are removed in the form of a wedge from the longitudinal center of the flitch. If the 2-inch flitch contains only one stud, it is ripped from the longitudinal center of the flitch with taper divided equally between the edgings. The 4-inch cant is ripped into studs parallel to and immediately adjacent to both bark edges. Thus taper and excess wood are removed from the juvenile core area rather than immediately under the bark as in Method III. Often a large part of the juvenile core is thus removed, and the percentage of core wood in all the studs from a given log is reduced. Like Method III, Method IV should yield a higher percentage of studs with balanced growth stresses than Methods I and II. Studs from 2-inch side flitches wide enough to yield two 2 by 4's tend to have less cross grain than similar studs developed in Method III. Eccentricity and compression wood.--most logs are eccentric, that is the pith is not in the exact geometric center. Generally, the more eccentric logs contain moderate-to-heavy concentrations of compression wood on the side with the longer radius (fig. 2). Positioning the log with the compression wood concentrated in the vertical or the horizontal plane as shown in figures 2 and 3 should result in studs with differences in warp tendencies. With vertical positioning the compression wood would tend to occur centrally with reference to the width of some studs (fig. 2), and with horizontal positioning the compression wood would tend to be concentrated on one edge or face of some studs (fig. 3). Because the effect of eccentricity may be independent of compression wood, the position of the eccentricity of all logs was controlled for each sawing method. An equal number of the logs in each variable class were sawed with the eccentricity vertical and horizontal. This might be considered. as doubling the number of sawing methods, but the author prefers to consider it as a separate controlled variable. Logs were also selected for presence or absence of compression wood visible on the log ends. Log position. --Because previous research (5, 9) indicated that warp potential of lumber cut from butt logs differs from that cut from upper logs, log position in the tree was also selected as a Figure 2.--Cross section with sawing Method III design applied to 10-inch log; cross-sectional eccentricity in vertical position. Compression wood is shown especially in studs 1 and 3. M Figure 3.--Cross section with sawing Method III design applied to a IO-inch log; cross-sectional eccentricity is in horizontal position. M

8 variable. In the sawmill it is usually possible to identify the butt logs but not possible to separate the upper logs further. Thus in this investigation all upper logs were combined into a single log position group. For substudies 1 and 2, equal numbers of butt and upper logs were selected. Only butt logs were used in substudy 3. The butt logs were defined as the first 8-foot log above the stump, and the upper logs as those from any position above the butt logs. Procedure Based on the variability noted in the previous study on loblolly pine (5), the log requirements were statistically predicted for the three substudies; 480, substudy 1; 180, substudy 2; and 240, substudy 3. Through the cooperation of the Intermountain Forest and Range Experiment Station, arrangements were made to select the logs from the stockpile of tree-length logs in the yard of the Elk Studs Company of West Yellowstone, Montana. These logs came from sales from the Gallatin and the Targhee National Forests. Most of the butt logs with the 6-inch requirement were selected from the Erickson pulpwood loading yard also located in West Yellowstone. Logs for this investigation were selected and bucked by personnel from the Forest Products Laboratory and the Intermountain Forest and Range Experiment Station. When the logs were selected there were no standard grading rules for lodgepole pine logs. Therefore the criteria were the same as used in the previous investigation (5) on loblolly pine. Thus all the logs chosen had the following general specifications: 1. Sweep of no more than 1 inch in the length of the log (8-1/4 to 8-1/2 feet). 2. No unsound knots. (An unsound knot is any visible branch, stub, or socket that contains either advanced decay extending to the log heart or any hole larger than 1/4-inch penetrating more than 2 inches (4).) 3. No bad knots. (Any visible knot larger than 1/6 of the log diameter (4).) 4. No evidence of decay, shake, or fire scars. For each log selected, the following information was recorded: Minimum and maximum d.i.b. to nearest 0.1 inch on both ends of the log, compression-wood class (compression or normal wood), log position class (butt or upper), and log diameter class (6, 8, 10, or 12 in.) based on minimum small end d.i.b. A metal tag bearing the log number was nailed to each end of each log. Logs were assigned a sawing method and an orientation-for-eccentricity class at random. These logs, with some additional logs to be used in perfecting the stenciling and sawing techniques, were shipped to the Laboratory in October If the weather warmed above the freezing point and there was danger of the logs drying out, they were kept under water-spray storage until sawed. Justbefore sawing, a thin cross-sectional slice was sawed from the upper end (usually smallest log end) to provide a new, clear surface. It was possible to stencil the exact sawing pattern on the end of each log. Black paint applied from aerosol cans worked well (fig. 4); then, a code number was placed on the end of each stud to identify Figure 4.--StenciIing sawing pattern with an aerosol paint spray for center cant on end of log. M the position of the stud in the log (fig. 5). A card indicating the log number was then attached, and the stenciled end of each log was photographed (figs. 2 and 3) for a permanent record. The logs were sawed on the Laboratory s sawmill. Although completely instrumented for research purposes, this sawmill is a conventional medium-weight circular headsaw type with a setter-controlled, hand-operated-blocks type carriage. The mill installation is equipped with a sawline indicator of the light projection type, which made it possible to position precisely the FPL 102 6

9 Figure 5.--Numbering system used to identify studs and their positions in a log. M opening saw cut on any face to coincide with the stenciled sawing pattern. Logs were placed on the carriage and securely dogged to the three headblocks with the stenciled end toward the saw. They were then sawed at a feed rate of about 1/8 inch per tooth, which compares favorably with fast commercial millproduction rates. Target size of the nominal 2 by 4 studs was 1-7/8 by 4 inches. Immediately after each stud was sawed, it was renumbered on the 4-inch face with both the log and the stud number. The studs were solid stacked until there was a sufficient number for a kiln load. They were covered with polyethylene plastic sheeting at all times to prevent drying. When a kiln load of studs had accumulated, studs were end trimmed to precisely 96 inches. Experience in the loblolly study (5) had shown that the rough edges and corners of the studs that resulted from sawing sometimes made precise measurement of deflection difficult--especially after kiln drying when the fuzz on the studs had stiffened. For this reason all the studs were planed lightly to 1-13/16 by 3-15/16 inches; this was sufficient to remove the fuzz and almost all saw marks. Each stud was then placed in a box that was constructed to provide the means to measure length and deflection. In this operation the length of each stud could be measured to the nearest 1/32 inch (fig. 6). The four faces of each stud were then numbered 7

10 Figure 6.--Measuring the length of a stud; metal panel immediately back of stud is graduated in increments of 1/32 inch. M Figure 8.--Measuring bow in a stud. M Figure 7.--Measuring crook in a stud; amount Figure 9.--Measuring twist in a stud; end of deflection is read on a tapered wedge of stud not shown is held flat to the gage. M table. M FPL 102 8

11 to indicate the position of each face in relation to the center of the log. The amount of crook and bow (figs. 7 and 8) was represented by another code number. Combinations of the two numbers indicated the amount and direction of the deflection. Twist was measured (fig. 9) and recorded to show degree and direction of rotation. After being measured, the 3,349 studs were dried in six separate kiln runs in the Laboratory's kilns to an average moisture content of 12 percent. After kiln drying, the studs were allowed to cool for 1 day and were then remeasured for length, crook, bow, and twist. The length of the studs that showed crook and bow was defined as the length of the chord connecting the ends on the concave face. This is a measure of the maximum usable length of the stud if it were cut squarely and parallel on both ends. Moisture content measurements were also taken at this time with an electric resistancetype moisture meter. When all kiln-dried studs had been remeasured, they were planed on four sides in a standard molder to finished dimensions (1-5/8 by 3-5/8 in.). After planing, they were again measured for moisture content, length, crook, bow, and twist. One hundred sixty-five, or 5 percent, of the studs were randomly selected for the 90-day storage test; they were carefully piled into six packages. Each package was strapped with three bands of 3/4- by inch steel. The banding was pulled as tightly as possible with commercial strapping tools to restrain a large part of the crook, bow, and twist. The packages were then placed in a dry open storage shed and allowed to stand 90 days to simulate conditions frequently encountered in the normal merchandising of lumber. After the 90-day storage period, the straps were removed from the studs, and the studs were allowed to stand for 2 days to permit stresses within them to reach a state of equilibrium. Then the studs were remeasured for moisture content, crook, bow, twist, and length. Analysis of Data All data were punched on cards, and analyses were made on the Laboratory's IBM 1620 computer. In most studs two criteria were used to evaluate the relationships of the variables to the amount of warp: The percentage of studs that met the No. 1 stud-warp limitations of the Western Wood Products Association (table 1); and the mean crook, bow, or twist for each treatment or for a combination of independent variables. Both evaluations were subjected to analysis of variance tests for statistical significance. Table 1.--Warp limitations for stud grades 1 1 According to Supplement 2, 1965 Standard Grading Rules, effective Oct. 15, 1966, Western Wood Products Association. Current rules issued July 1, 1960, are similar if grade 1 is changed to 2 and Better, grade 2 to grade 3, and grade 3 to grade 4 and specific requirements for bow are dropped. 2 More than allowed in No. 3 for crook, bow, or twist. Substudy 1 In this study, examination was made of the relationship of sawing Methods III and IV, of butt and upper logs, of 6-, 8-, 10-, and 12-inch diameter classes, and of horizontal and vertical eccentricities to the subsequent warp in dry studs. Yields (by percent) of No. 1 studs are listed in tables 2 through 5. Sawing Methods III and IV.--SawingMethod IV tended to give slightly higher percentages of No. 1 studs from both butts and uppers combined than Method III when considering only crook (table 5, 91.1 vs pct.) or all warp (table 5, 75.9 vs pct.). The term all warp used here means the grade of the stud determined by considering the warp aspect that relegates a stud to the lowest grade. Thus a stud with No. 1 crook, No. 2 bow, and No. 3 twist is No. 3 all warp. It is equivalent to its actual grade established by the warp limitation in the grading rules. The slight superiority of Method IV, although reasonably consistent throughout the work, is not sufficient to be of statistical significance. It follows similar findings for loblolly pine (5); yet in the author's opinion this slight superiority constitutes a real difference. 9

12 Differences in sawing Methods III and IV appeared to have no influence on the percentage of studs meeting the No. 1 warp requirements for twist and bow. Differences in yields when measured by percent of studs meeting No. 1 bow limitations are very insensitive because the bow limitation for No. 1 is very large (table 1, 24/32 in.). When crook, bow, and twist were evaluated by their mean deflection for a specific combination of variables, no significant differences were indicated between sawing Methods III and IV. Log position (butts vs. uppers).--log position had a significant influence on warp in almost all studs; this is in complete agreement with earlier findings on loblolly pine (5). For studs that met No. 1 stud-grade limitation for crook, superiority of upper logs (table 2, 95.0 pct.) over butt logs (table 2, 85.7 pct.) was highly significant. When twist is considered, the results were the opposite. The superiority of butts (table 4, 90.5 pct.) over uppers (table 4, 77.9 pct.) was highly significant. This also agrees with- observations on loblolly pine (5). In evaluating bow, the superiority of uppers (table 3, 99.4 pct.) over butts (table 3, 97.7 pct.) approached significance at the 0.05 level. The superiority of butts for twist of uppers for crook tended to nullify each other when all-warp yields were considered. Butts at 76.3 percent (table 5) are numerically superior to uppers at 73.8 percent (table 5), but the difference is not statistically significant. The results when mean deflection is used as the criterion for evaluating warp are shown in figure 10. When crook is considered, the difference between studs from butt logs, 4.5/32 inch, and studs from upper logs, 2.5/32 inch is highly significant. The same relationship is noted lor bow in which studs from butt logs show an average deflection of 7.2/32 inch and uppers only 3.6/32 inch. The studs from upper logs show substantially more twist (5.9/32 in.) than those from butt logs (4.3/32 in.). This difference, as for crook and bow, is highly significant. Diameter.--Based on the percentage of studs that met warp limitations for No. 1 stud-grade, diameter seemed to have no effect on crook or bow in studs. A significant relationship was found between increasing diameter and decreasing twist, Using the yields shown in table 4 and determining Table 2.--Percent yields of grade 1 studs based on Western Wood Products Association warp Iimits for crook 1 Averages in "AII" and "Combined" columns are not necessarily the arithmetic mean of the other tabular values because they are weighted by the actual number of studs in each class. FPL

13 Table 3.--Percent yields of grade 1 studs based on Western Wood Products Association warp limits for bow Table 4.--Percent yields of grade 1 studs based on Western Wood Products Association warp limits for twist 11

14 Table 5.--Percent yields of grade 1 studs based on Western Wood Products Association combined warp limits for all warp (crook, bow, and twist) the yields by diameter for all logs sawed by Methods III and IV combined, the diameter-yield relations are 79.3, 74.5, 81.8, and 91.0 percent for 6-, 8-, 10-, and 12-inch diameters. There is some question that the indicated statistical significance of this relationship may be caused by the high value for the 12-inch logs in combinations with the low value for the 8-inch class because both 6- and 10-inch diameter-yield relations are about the same. If all warp is considered, diameter did not account for any differences in yield. The decrease in mean twist with increase in diameter for all logs shown in table 6 (6 in., 5.4/32 in.; 8 in., 6.4/32 in.; 10 in., 5.0/32 in.; and 12 in., 3-4/32 in.) was highly significant for all studs. The relation of diameter to crook was significant only in upper logs (6 in., 1.9/32 in.; 8 in., 2.5/32 in.; 10 in., 2.2/32 in.; and 12 in., 3.4/32 in.) indicative of an interaction with log position. The relation of diameter to bow was highly significant only in butt logs (6 in., 12.9/32 in.; 8 in., 5.8/32 in.; 10 in., 5.6/32 in.; and 12 in., 4.5/32 in.) indicative of another interaction with log position, Eccentricity position.--for the percentage of studs that met No. 1 stud limitations for crook, both butt and upper logs from all diameters combined sawed by Methods III and IV yielded from 1.0 percent (upper logs, Method IV) to 7.7 percent (butt logs, Method III) more top grade if sawed with the long radius oriented vertically (table 2). This difference was reasonably consistent throughout the work; however, it is not statistically significant. There was no apparent relationship between the rotational position of the eccentricity and the percentage of No. 1 studs based on bow, twist, or all-warp limitations. A highly significant interaction between the position of the eccentricity and the log position was noted when measured by mean bow. From table 7, butt logs sawed in the vertical position had lower bow (6.9/32 in,) than those sawed in the horizontal position (7.6/32 in.). In upper logs, horizontal positioning gave lower yields than did the vertical (3.2/32 vs. 3.9/32 in.). There was no significant relationship between eccentricity and crook or twist although vertical positioning of the eccentricity gave lower mean values for crook in almost all cases. FPL

15 Substudy 2 Figure 10.--Effect of log position on crook, bow, and twist for mean warp of all studs. M Examined in substudy 2 was the relationship of sawing Method I to sawing Methods III and IV in butt and upper logs of horizontal and vertical eccentricities in only the 12-inch-diameter class. Yields in percent of all stud grades are shown in table 8 and figure 11. Yields for crook, bow, and twist are given in mean warp in tables 9 and 10. Sawing Methods I vs. III and IV.--The results of this comparison are shown in table 8 for percent of studs that met the all-warp limitations for the grades of the Western Wood Products Association. Sawing Method IV was significantly better than sawing Method I for percent of studs meeting grade 1 specifications for all warp (83.5 vs pct.). Sawing Method III was also better, but not statistically (79.3 vs pct.). However, the combined sawing Methods III and IV were significantly better than sawing Method I. This observation is in general agreement with earlier findings for loblolly pine (5); however, the difference was substantially greater for loblolly pine. There is some indication that sawing Method IV is better than sawing Method III if 12-inch logs are sawed although this was not significant statistically (83.5 vs pct.). If sawing methods are compared based on mean crook of the studs, much the same trends are evident. From table 9, mean crook for sawing Method I is 6.6/32 inch compared with 3.9/32 inch for Method III and 3.5/32 inch for Method IV; Table 6.--Mean warp in increments of 1/32 inch by diameter classes of butt and upper logs1 13

16 Table 7.--Mean warp in increments of 1/32 inch by eccentricity classes for butt and upper logs Table 6.--Percent yields from 12-inch-diameter logs according to Western Wood Products stud grades1 FPL

17 Table 9.--Mean warp in increments of 1/32 inch for 12- inch Iogs1 Table 10.--Mean warp in increments of 1/32 inch in studs from 12-inch logs1 15

18 Figure 11.--Percent yields for sawing methods by stud grades of WWPA (Western Wood Products Association) of 12-inch logs (butts and uppers) in substudy 2. M these differences are highly significant. Log position was also significant when evaluated by mean crook with uppers superior to butts by 3.7/32 to 5.6/32 inch. No significant difference was found between mean bow for sawing Methods I (4.5/32 in.), III ( in.), and IV (4.1/32 in.). Again, as in substudy 1, upper logs produced studs with significantly less bow (3.6/32 in.) than butt logs (4.8/32 in.). If mean twist is considered, no significant difference was noted between sawing Methods I (3.9/32 in.), III (3.7/32 in.), and IV (3.2/32 in.). The superiority of studs from butt logs (2.9/32 in.) over studs from upper logs ( in.) was highly significant. Studs from logs with vertical eccentricity had significantly less twist (3.2/32 in.) than those with horizontal eccentricity (4.1/32 in.) as shown in table 10. Substudy 3 Examined in this study was the relationship of the presence or absence of visually evident compression wood on the log ends, of 8- and 10-inch-diameter butt logs, of sawing Methods III and IV, and of horizontal and vertical eccentricities to the subsequent warp in dry studs. Yields for stud grades are shown in table 11. Compression wood. --The results for percent of studs meeting No. 1 stud-grade requirements show no significant differences between studs from logs with or without visibly evident compression wood. However, it is interesting to note in table 11 that logs with no compression wood had a higher yield in all categories except the 10-inch logs sawed in the vertical position by Method III. No effects of compression wood were noted FPL

19 Table 11.--Percent yield from 8- and 10-inch diameter butt logs according to grades of the Western Wood Products Association1 when results were evaluated for mean crook, Relationship of Position of bow, and twist. Stud in Log to Warp Other variables.--based on the percent of studs meeting grade 1 requirements, sawing Table 13 shows the percentage of studs from methods, eccentricity, and diameter had no sig- the inner and the outer areas of the log within nificant effect on the warp of studs. the limits of No. 1 studs for crook, bow, twist, With the exception of crook, this was also true and all warp by diameter and by sawing method. when mean warp was used as the criterion. When Figures 12 through 19 show the yields of grade 1 crook alone was considered, the sawing method- studs from each stud position for each form of eccentricity interaction was significant. From warp in both butt and upper logs. table 12, sawing Method III produced lower mean For sawing methods III and IV, stud numbers crook results with vertical eccentricity than with considered inner were as follows: 8-inch, No. 11; horizontal eccentricity (3.7/32 vs. 5.8/32 in.). 10-inch, Nos. 1 and 2; 12-inch, Nos. 11, 12, and However, when sawing Method IV was used, 13 (fig. 5). Studs Nos. 71, 72, 81, and 82 were horizontal eccentricity produced less crook than considered inner for sawing Method I. All other vertical (4.2/32 vs. 4.9/32 in.). studs were considered outer. A separation of the 17

20 Table 12.--Mean warp in increments of 1/32 inch for vertical and horizontal eccentricity Combining all aspects and considering the allwarp limitations, yields of No. 1 studs from the outer area (76 pct.) exceeded by about 8 percent the yields from the inner area (68 pct.); this difference was statistically significant. For sawing Method I, yields from the outer area were substantially higher when measured for crook (86 vs. 67 pct. for butts, 97 vs. 84 pct. for uppers). Similar differences also were found when all warp was considered (78 vs. 49 pct. for butts and 88 vs. 77 pct. for uppers). Very little variation was noted when bow and twist were used as a measure. Recoveries of Studs in All Grades studs from 6-inch logs into inner and outer categories was not possible because only two studs were involved, and both were in approximately the same relative position in the log. For sawing Methods III and IV in both butt and upper logs, the increased yield of No. 1 studs based on crook from the outer area compared to the inner area was highly significant. The superiority of outer area studs was more pronounced in butt logs than in upper logs. From table 13, 90 percent of the outer and 74 percent of the inner butt studs are grade 1, whereas the comparable figures for upper logs are 97 percent and 91 percent. This, too, agrees with similar findings on loblolly pine (5). When bow is considered, results for Methods III and IV indicate the increased yield of No. 1 studs from the outer area of both butt and upper logs to be highly significant. There was, however, less difference in the upper logs (100 vs. 98 pct.) than with butt logs (100 vs. 96 pct.). There was also an interaction with sawing method; the difference in yield between inner and outer studs was smaller for sawing Method III than for Method IV. No. 1 stud yields based on twist limitations did not show a consistent tendency, and there was no significant relationship to stud position in log, log position, or sawing method. Yields of inner studs exceeded those for outer from butts by about 6 percent (93 vs. 87 pct.) and from upper logs, it was the reverse, the outers exceeded the inners by about 4 percent (78 vs. 74 pct.). Although studs are frequently marketed in combined grades such as No. 2 and Better (2 & Btr.) or No. 3 and Better (3 & Btr.), evaluation of yields from these combined grades does not show quality differences in the mix. Data are given in tables 14 and 15 for all grades sawed by Methods III and IV. Yields are further subdivided by log position and by eccentricity class. There is a tendency for combinations of variables that show high yields of grade 1 to show low yields of grade 3 with little change for grade 2. Usually the yield of Below Grade studs was so small that it was unimportant. Studs After 90-Day Storage Five percent (165) of all of the studs (randomly selected) were strapped and stored for 90 days in a dry, open shed. The results of this storage on stud length, crook, bow, twist, and moisture content is shown in table 16. There was a substantial tendency for the studs to elongate during their storage period. Sixtytwo percent increased in length, 12 percent b e c a m e shorter, and 26 percent remained unchanged. Most of the lengthening was confined to 2/32 inch or less although four studs increased 1/8-inch in length. The overall result of storing under restraint on crook, bow, and twist was minor. There was a slightly greater percentage of studs in which the crook and bow increased (31.4 and 39.1 pct.) than decreased (27.7 and 34.3 pct.). Twist showed a somewhat greater and reversed tendency to decrease rather than increase during storage (27.1 vs pct.). FPL

21 Table 13.--Percent yield of grade 1 studs for inner and outer stud positions1 19

22 Figure 12.--Effect of position of stud in log on amount of crook in butt logs. (Based on percentages of studs that met crook Iimitations for WWPA "stud" grade.) M FPL

23 Figure 13.--Effect of position of stud in log on amount of crook in upper logs. (Based on percentages of studs that met crook Iimitations for WWPA "stud" grade.) M

24 Figure 14.--Effect of position of stud in log on bow in butt logs. (Based on percentages of studs that met bow Iimitations for WWPA "stud" grade. ) M FPL

25 Figure 15.--Effect of position of stud in log on bow in upper logs. (Based on percentages of studs that met bow Iimitations for WWPA "stud" grade.) M

26 Figure 16.--Effect of position of stud in log on twist in butt logs. (Based on percentages of studs that met twist Iimitations for WWPA "stud" grade. ) M FPL

27 Figure 17.-Effect of position of stud in log on twist in upper logs. (Based on percentages of studs that met twist limitations for WWPA "stud" grade.) M FPL

28 Figure 18.--Effect of position of stud in log on warp in butt logs. (Based on percentages of studs that met ali warp Iimitations for WWPA "stud" grade.) M FPL

29 Figure 19.--Effect of position of stud in log on warp in upper logs. (Based on percentages of studs I-hat met ali warp Iimitations for WWPA "stud" grade.) M

30 Table 14.--Percent yield from butt logs Table 15.--Percent yield from upper according to stud grade of Western logs according to stud grades of Wood Products Association 1 Western Wood Products Association 1 28

31 Table 16.--Changes in moisture content, length, and amount of warp in 165 studs strapped during 90-day storage During the storage period the moisture content in 44.6 percent of the studs decreased, whereas it increased in 24.7 percent. In summation, the studs in storage tended to become drier and to elongate; croak and bow increased slightly, but twist was reduced. 29

32 Summary Investigated in this work is the relation of sawing method and other related variables to the warp of nominal 2-inch by 4-inch studs sawed from small lodgepole pine logs. Other variables included the position of the stud in the log, rotational position of the log cross-sectional eccentricity, log diameter, log position in the tree, and the presence of compression wood. A total of 680 logs were sawed into 3,349 studs on the sawmill at the Forest Products Laboratory; the studs were kiln dried to an average moisture content of 12 percent. After drying and planing, the studs were measured and evaluated to determine crook, bow, twist, moisture content, and length, Both percentage of studs meeting WWPA (Western Wood Products Association) stud grade 1 warp requirements and mean warp were used as evaluation criteria. Diameters used were 6-, 8-, 10-, and 12-inches inside bark on the small end of the log. Logs were classified as butts or uppers. Compressionwood logs were those with sufficient compression wood to be visible on one or both log ends. Logs were sawed according to two commercial methods and one experimental sawing method with each applied to an equal number of logs in which eccentricity (longest radius) was oriented horizontally and vertically. Both sawing methods, the Scragg (Method 111) and the FPL improved Scragg (Method IV) were markedly superior to the conventional method (I) both for percentage of grade 1 studs produced and for the lowering of the mean warp of the studs. Method IV was generally better than Method III in controlling crook, but the difference was not great. The yield of studs from butt logs was more prone to crook and bow but less prone to twist than it was from upper logs. On a grade-yield basis, diameter affected the yield only when twist was considered. The larger logs showed lower percentages of badly twisted studs. Mean warp further indicated that bow likewise decreased with increased diameter. In upper logs crook was lowest for the smaller logs. Placement of the eccentricity Vertically generally resulted in less warp and more grade 1 studs. No difference in yields of grade 1 studs or in their average warp could be attributed to the presence or absence of visually evident compression wood. Because the warping effect of compression wood especially if located along one edge or face of a stud can hardly be denied, the only rational conclusion is that compression wood is not easily identified visually on the log ends in lodgepole. Thus it is probable that a substantial part of the sample classified compression wood absent actually contained compression wood. Studs from the inner area showed consistently more crook and bow than those in the outer area near the bark, The position of the stud in the log did not seem to influence twist. One hundred sixty-five studs, a randomly selected 5 percent sample, were tightly strapped and stored in an open dry shed for 90 days. On remeasurement it was determined that the studs had become drier and longer and crook and bow had increased slightly but that twist had decreased. FPL

33 Recomendations to the Industry From this research the following suggestions are made to sawmills producing studs from lodgepole pine: 1. Use either the Scragg or the FPL improved Scragg sawing method. 2. If crook is a major problem, consider using the FPL improved Scragg method and saw only upper logs into studs. Butt logs could be better used in lumber with wider dimensions that has more inherent resistance to crook. 3. If twist is a major problem, it can be reduced by increasing the percentage of butt logs and by increasing the average diameter of the log mix. 4. It is probaldy worth the possible small additional cost to orient all logs to have the long radius vertical before reducing the logs to cants and flitches. 31

34 Literature Cited 1. Cockrell, R. A Some observations on density and shrinkage of ponderosa pine wood. Amer. Soc. Mech. Engin. Trans. 65: pp., illus Further observations on longitudinal shrinkage of softwoods. Forest Prod. Res. Soc. Proc Du Toit, A. J A study of the influence of compression wood on the warping of Pinus radiata D. Don. timber. S. African Forestry J. No. 44: Englerth, G. H., and Smith, W. R Guide for grading southernpine logs. Southeastern Forest Exp. Sta. Asheville, N.C. July. 5. Hallock, H Sawing to reduce warp of loblolly pine studs. U.S. Forest Serv. Res. Pap. FPL 51, 52 pp., illus. Forest Prod. Lab., Madison, Wis. 6. King, Woodrow W Cause of and remedy for warped southern yellow pine 2 x 4 s. Texas Forest Serv. Circ. No. 37, 19 pp., illus Alleviating bow and crook in southern yellow pine dimension, with chemicals. Forest Prod. J. IV(5): Kotok, E. S Tree characteristics influence 2 x 4 stud yield of lodgepole pine. U.S. Forest Serv. Res. Note INT-63, 8 pp., Intermountain Forest and Range Exp. Sta., Ogden, Utah. 9. Kloot, N. H., and Page, M. W A study of distortion in radiata pine scantlings. Commonwealth Sci. and Ind. Res. Organ. Div. Forest Prod. Tech. Rep. No. 7, Melbourne, Australia, 24 pp. 10. Paul, Benson H The juvenile core in conifers. Tappi 43(1): , and Sweet, C. V Quality control in manufacture of lumber from second growth. Rep. 1781, U.S. Forest Prod. Lab., Madison, Wis. 12. Pillow, M. Y Compression wood cause of bowing and twisting. Wood Working Ind. 8(5): U.S. Forest Products Laboratory Wood handbook, U.S. Dept. Agr., Agr. Handbook No. 72, 528 pp Longitudinal shrinkage of wood. Rep. 1093, U.S. Forest Prod. Lab., Madison, Wis. 15. Zohel, B., Webb, C., and Henson, F Core or juvenile wood of loblolly and slash pine trees. Tappi 42(5): FPL