Ai u - s - ' ' REGION 8 L 7— : ; — t. MAY 1 2 1937 U. S. Department of Agriculture, Forest Service FOREST PRODUCTS LABORATORY In cooperation with the University of Wisconsin MADISON, WISCONSIN £-[0*3 THE LONGITUDINAL SHRINKAGE OF WOOD By ARTHUR KOEHLER In Charge, Section of Silvicultural Relations Published in Transactions of the Wood Industries Division of the AMERICAN SOCIETY OF MECHANICAL ENGINEERS April, 1931 Digitized by the Internet Archive in 2013 http://archive.org/details/longhriOOunit LOiTGITUDIlTAL SHRINKAGE 0? V/OOJji By ARTHUR KOEKLSR^ Longitudinal shrinkage of wood ordinarily is so small as to "be considered negligible, and no allowance is made for it for most uses.. There are uses, however, for which even the small longitudinal shrinkage of normal wood is an important factor, and others in which the shrinkage along the grain of abnormal wood is so large as to be a serious handicap. This paper gives information on the shrinkage of wood along the grain about which very little has been published. Shrinkage along the grain of wood may be objectionable under the following conditions; 1. When it causes undue shortening in length, particularly in posts, so as to throw the adjoining members into serious disalignment. As a rule, no allowance is made for longitudinal shrinkage as is usually done for transverse shrinkage, and consequently when it does occur to an appreciable degree it often causes trouble. 2. When it causes exceptionally wide opening of butt joints. 3» When it causes bowing or cross breaks in pieces which shrink unevenly longitudinally. 4« When it weakens wood in longitudinal shear on account of stresses set up as a result of variations in shrinkage potential of different layers in timber or other members. 5» When it causes slivers to loosen and raise up from the face of a board. —Presented at the Fifth national Wood Industries Meeting, Uew York, N. Y., October 16 and 17; 1930; of The American Society of Mechanical Engineers. 2 In charge of Section of Silvicultural Relations, Forest Products Laboratory, Forest Service, U. S. Department of Agriculture, maintained in cooperation with the University of Wisconsin. Mr. Koehler has been engaged in the study of the cellular structure of wood for identification purposes and to explain the variations in the behavior of wood since 1911* R1093 Variations in Shrinkage The data here presented are the result of an initial investi- gation by the Forest Products Laboratory into causes for the variation in the shrinkage of wood. There are large variations in the shrinkages of the different species of wood, in the individual pieces of the same species, and in different directions on the same piece. These variations can be accounted for primarily by differences in the cellular structure. The longitudinal shrinkage of wood normally ranges from 0.1 to 0.2 of 1 percent. Unless otherwise stated all shrinkage values given in this paper represent the total shrinkage from the green or soaked to the oven-dry condition, that is, a condition of zero moisture content. In drying to an average air-dry condition of about 12 percent, the shrinkage would be about one-half as much. Cases have been observed in which the shrinkage along the grain was abnormally large. The greatest longitudinal shrinkage so far recorded at the Forest Products Laboratory is 5»?8 percent in wood from the lower side of a limb of western yellow pine* This is more than the average transverse shrinkage for the species. It is equivalent to a shortening of 11.1 inches in a l6-foot board. On the other extreme, slight elongation has been observed in certain woods in drying. Figure 1 gives shrinkage values for 3^ different types of wood at three different percentages of moisture content; namely, about 12 percent, about 7 percent, and at percent. Shrinkage is assumed to have begun at a moisture content of 24 percent, which is taken as an average fiber-saturation point. Different types of wood were selected so as to give as wide a variation in shrinkage as possible; consequently the values given in Figure 1 are not intended to represent averages for any species. Conditions Under T Thich Excessive Longitudinal Shrinkages Occur No sharp limits can be set above which longitudinal shrinkage in wood may be called excessive. A total uniform shrinkage potential of 0.3 of 1 percent, however, is near the maximum permissible for most uses. It would cause a shortening of O.^S inches in a lb-foot board if oven dried. For numerous uses, 0.2 of 1 percent would be the maximum permissible. Fortunately most wood falls below these limits* If the longitudinal shrinkage is not uniform from side to side, an even smaller shrinkage may be objectionable on account of the bowing which it may produce. Shrinkage beyond these. limits has been found to occur in the following types of wood: R1093 -2- 1. Compression wood . --Compression \/ood is a hard, heavy, brittle type of wood formed on the lower side of branches and leaning trunks of coniferous trees. It is characterized by rather wide annual rings containing wide summerwood that is not so dark and hornlike as in normal wood. Figure 2 shows a cross section of a spruce log in which compression wood is well developed. Within any annual ring compression wood is confined to one side of the pith, that is, it does not extend continuously around the trunk. It may be continuous from the pith to the bark, or it may occur in intermittent layers. In some annual rings it may be in one direction from the pith and in others in a different direction. Therefore, it frequently happens that boards or dimension stock have streaks of compression wood in them. Compression wood, having a greater tendency to shrink longitudinally than normal wood, usually causes a bowing of lumber in which it is unevenly distributed. (See fig. 3«) Occasionally streaks of compression wood that adjoin normal wood will pull themselves apart so as to form cracks across the grain. Compression wood is a source of considerable trouble in the drying and use of softwood lumber, principally on account of the bowing which it produces. Sometimes the direct shortening causes trouble. For example, butt joints in redwood and cypress siding occasionally open up on account of the longitudinal shrinkage of compression wood in the siding. If the siding is well-dried before it is nailed down, however, there will be no occasion for any further shrinkage to take place. No sharp line of distinction can be drawn between compression wood and normal wood; the two grade into each other imperceptably. Therefore, no fixed shrinkage value can be ascribed to compression wood. The maximum longitudinal shrinkage so far found in compression wood has already been given for wood from a limb of western yellow pine. The maximum shrinkage so far determined for compression wood from a tree trunk is 5*42 percent in a southern yellow pine. More often the shrinkage of compression wood is less than 1 percent but over 0.3 of 1 percent. 2. Abnormally light -weight wood . — Wood below the average weight for a species shrinks more along the grain than heavier wood of the same species. This is shown in Figure 1 for high- and low-density wood of longleaf pine (1 and 2), tupelo gum (22 and 23); ash (25 and 26), and for red oak of low density which shrunk over 0.8 of 1 percent. Figure 4, which shows the shrinkage of redwood of different specific gravities, clearly indicates a general trend toward a reduction in shrinkage with an increase in specific gravity., All of the specimens represented in Figure 4 above the 0*4-00 percent line and about one-half of those between the 0.300 and 0.U00 percent lines consisted of compression wood. With this fact in mind, the downward trend of the remaining points which represented normal wood almost entirely, is more apparent. Balsa, on the other hand, which normally is a very light wood, did not shrink so much as most of the other specimens that were light in weight. (See fig. 1.) It is only wood which is considerably below the average weight for the species that shrinks excessively along the grain. R1093 -3- This peculiar relationship is due to the; fact that in normal- weight wood most of the fibrous cells, or wood fibers, are constructed so that they do not shorten much in drying, and since the wood fibers predominate, as a rule, they control the shrinkage of wood as a whole. In abnormally light-weight wood, however, either the wood fibers have a different type of structure or they are so scarce that the other kinds of cells predominate, particularly those that make up the pores, wiJ.ch have a high shrinkage potential. The relationship between longitudinal shrinkage and density is just the opposite of what it is for transverse shrinkage and density; that is, in general the heavier the wood the more it shrinks across the grain. The relationship between density and shrinkage, whether longitudinal or transverse, is caused by the fact that the woods of low and high densities have different cellular stractures, and not by the fact that there is a difference in the amount of wood substance present. 3» Sprin,;wood ."-In many species of wood each annual growth layer is differentiated into an inner, softer layer of low density, the springwood, and an outer, harder layer of high density wood, the summenvood. The springwood invariably shrinks more along the grain than the summerwood. This is easily demonstrated oy whittling out a piece of wood only one annual ring in thickness and observing the way it bows lengthwise of the grain as it dries. The greater longitudinal shrinkage of springwood as compared with summerwood is shown in Figure 1 for loblolly pine (3 and *+) , southern yellow pine (5 and 6), abnormal southern yellow pine (7 and 8), Douglas fir (10 and 11), and redwood (13 and lk) , Figure k shows that the specimens of redwood of lowest specific gravity, which consisted entirely of spring- wood, shrank considerably, whereas those of highest specific gravity, which consisted entirely of suramerwood, shrank very little. Figares 1 and h show that some pieces even elongated in drying. Such elongated pieces invariably consisted of summerwood only. Within a piece of wood several annual rings in width, the springvvood, of course, cannot shrink more than the summerwood since the two are so closely and alternately bound together. The difference in shrinkage potential, however, undoubtedly sets up stresses which may possibly produce shear failures either directly or when supplemented by stresses due to external loads. In flat-grain lumber the greater longitudinal shrinkage of springwood frequently causes slivers to loosen and curve away from the surface, especially on the pith side. This condition frequently is exasperating in flat-grain flooring. It can be largely avoided by dressing pattern stock so that the face side is on the bark side of the piece. In that event, the summerwood of each annual ring is toward the outside, and consequently the tendency of the annual rings to curve up is eliminated. RIO 93 -li- Rotary veneer occasionally happens to be cut so as to be one annual ring in thickness. Under such circumstances, any difference in shrinkage potential of springwood and summerwood causes local curvatures. h. Wood near the pith of the tree . --Boards and planks of the southern and western yellow oines, and occasionally other species, which happen to be sawed lengthwise through the pith, frequently sho?/ cross breaks extending for a short distance at right angles to the pith. (See fig, 5«) Such breaks are due to the greater longitudinal shrinkage of the wood near the pith. For the same reason, narrow strips cut out so that the pith runs along one edge will bow as they dry. On the other hand, it lias been reported that the wood on the bark side of some hardwood shrinks more than the wood away from the bark, and thereby causes bowing. The explanation probably is that the density of the wood in such pieces decreased toward the outside of the tree, as frequently is the case, and as previously stated, low density wood shrinks more longitudinally than that of high density. 5» Wood in very fast growing conifers . — In loblolly pine and redwood relatively high longitudinal shrinkage has been found in second- growth trees with very wide annual rings containing summerwood which is not so dark and hard as in normal wood. Such wood resembles compression wood in some respects, but differs from it in extending entirely around the tree trunk and in some microscopic features. Wood of this type was found mostly in the butt portion of the trunk. Probably it also occurs in other species than the ones mentioned. 6. Cross-grained wood . — In any piece of wood in which the grain does not run parallel with the long axis, there may be excessive apparent longitudinal shrinkage due to a transverse shrinkage component being affective along the length of the piece. If the grain ran at an angle of ^5 degrees across a board it would shrink as much on a percentage basis along its length as along its width. Wood with spiral, diagonal, inter- locked, wavy, or curly grain may have excessive apparent longitudinal shrinkage on account of the presence of cross grain. A knot along one edge of a narrow strip will cause bowing of the strip due to the cross grain which it produces. 7« Long pieces of wood . — Although a long piece of wood should not shrink any more than a short one of the same kind on a percentage basis, the actual shrinkage is proportional to the length. The greatest shortening effect therefore, would be expected in long boards, planks, timbers, or the like. The present tendency toward increased use of short lengths will, therefore, help to break up the longitudinal shrinkage into smaller units, which will be particularly advantageous in keeping down the width of openings in butt joints in siding and flooring. R1053 -5- Relation of Longit udina 1 5. a' inhale to Moisture Cent' .. t No attempt has been made by the author to determine the moisture content at which wood begins to shrink longitudinally in drying. It is hardly conceivable that it would be any different from that at which transverse shrinkage occurs; namely, the fiber-saturation point, which averages about 2*4 percent moisture. Although most of the points representing shrinkage for different percentages of moisture content in Figure 1 are not in a straight line or characteristic type of curve for each kind of wood, they agree fairly well in pointing to about 2.h percent as the moisture content at which shrinkage began. Among the specimens of low shrinkage there is a slight tendency to curvature in the shrinkage lines, indicating that the shrinkage rate increases as the moisture content decreases. Some of the specimens showed no shrinkage or even elongated in the first and second stages of drying, but when all moisture was removed they v/ere shorter than when wet. Longitudinal shrinkage, like transverse shrinkage, depends on the change in moisture content below the fiber-saturation point. If wood is dried so that it will not change a great deal in moisture content when used, then it will not shrink appreciably along the grain, no matter how abnormal its shrinkage potential may be. For uses in which the wood becomes alternately wet and dry, previous drying, of course, does not prevent subsequent changes in dimensions with changes in moisture content. As previously pointed out, the uses discussed here in which longitudinal shrinkage becomes an important factor are far in the minority; for most uses the longitudinal shrinkage of wood is so small as to be considered negligible. Future Fork Further investigations are under way at the Forest Products Laboratory to determine how the cellular structure of wood affects its shrinkage. It is believed that information on the cause of variations in the shrinkage of wood will be of value in diagnosing and explaining peculiarities in the behavior of wood, and in learning how to recognize wood which has an exceptionally large shrinkage potential, so that diffi- culties that might result therefrom may be guarded against. It is also believed that an under standing of the mechanism of wood shrinkage will be of value in developing treatments for reducing the shrinking and swelling R1093 -S- of wood in use. It has already "been discovered that the pitch of the spirally wound cellulose fibrils of which the cell wall is composed is one of the controlling elements in longitudinal shrinkage; the less the pitch the greater the shrinkage in the length of the cell. Theoretically, the reverse should hold for transverse shrinkage, and it was found to do so in many cases. For example, compression wood shrinks less than normal wood, and springwood shrinks less than summerwood across the grain. Evidently other factors also come into play in transverse shrinkage, the importance of which should be determined by the further investigations now under way at the Forest Products Laboratory. R1093 -7- O 13 -P a «H -P o o «H O u a > ^ 5 Figure 2. — Cross section of spruce log showing compression wood — Zr0feSS6 p the dark, wide-ringed portion. VI • o »d o o & 01 d CD O CO -P [0 S3 CD 0) «h a o o •H i o3 U CO d) » o rH O 3* ,0 O o -H «H CO CO *j c * * ^?5 X X ' . Q. U. -i io io ^ 1 1 1 . 4 X X K X • • i • • x X ***> o • 9 • m • • • * • • • ♦ • • • • • • • h* i— • • • • • ► * • « **•: • ••* • * • • X «.~ • • . * • • • • • t • • • i • • • • . ■-'.: •••• Q • • •• .. • ; • i:« V* • • • ". • . « > • • • • • • • • • 4 q # .i 4 4 4 4 <* 4 r 4 Q | I i < 1 : 1 i a r 5 * V 1 i \ \ >5j cc sfl -P •H i o •H «H •H O & w 0) -p o -p -d o o •S 8 o CO «H O a o I P4 H1DN31 N33HD JO 1/V3J d3d — 30VMNIdHS It/NlonilSNOl u CD CD 1 -P •H ft (-1 n3 0) •H T* ft o o Is * o r-H »