IAWA Bulletin n.s., Vol. 10 (4), 1989: 364-373

WOOD IDENTIFICATION OF COMMERCIALLY IMPORTANT NORTH AMERICAN SPECIES OF BIRCH (BETULA)

by Regis B. Miller and Eric Cahow USDA Forest Service. Forest Products Laboratory, One Gifford Pinchot Drive, Madison, WI 53705-2398, U. S. A.

Summary The wood anatomy was studied of the key, however, was not incorporated into text­ four commercially important North American books or later cited by authors. Panshin and species of birch: Betula alleghaniensis, B. De Zeeuw (1970, 1980) and Panchin et al. lenta, B. nigra, and B. papyrifera. Although (1964) grouped five North American tree the wood from these species is fairly homo­ species of birch and made no mention of geneous. it appears that B. papyrifera can be diagnostic features for species separation. separated from B. nigra; B. alleghaniensis Core et al. (1976, 1979) stated ‘the various and B. lenta, though indistinguishable from birches cannot be separated.’ To determine if each other, can be separated from both B. reliable diagnostic features are present to papyrifera and B. nigra. The diagnostic fea­ separate species or species groups, we stud­ tures are as follows: ray width, ray and ray ied the wood anatomy of the following four cell shape as viewed on the tangential longi­ commercially important North American spe­ tudinal section, average number of bars per cies of birch: yellow birch, B. alleghaniensis perforation plate, and contents of axial paren­ Britton (= B. lutea Michx. f.); sweet birch, chyma cells. In addition to these species, two B. lenta L.; river birch, B. nigra L.; and pa­ commercially important European species (B. per birch, B. papyrifera Marsh. For com­ pendula and B. pubescens) and two addi­ parison, we examined the two commercially tional North American tree species (B. occi­ important European birches, B. pendula Roth dentalis and B. populifolia) were studied. A (= B. verrucosa Ehrh.) and B. pubescens dichotomous key to all eight species is pre­ Ehrh., and two additional North American sented. tree species, e.g. water birch, B. occidentalis Key word: Wood anatomy, wood identifica­ Hook. (= B. fontinalis Sarg.) and gray birch, tion, Betula, birch, Betulaceae. B. populifolia Marsh. Virginia roundleaf birch [B. uber (Ashe) Fern.], the only other Introduction North American tree species of Betula, was In the forest products industry and in oth­ not examined, because it is extremely rare er identification situations, it is often de­ and specimens were not available. sirable to distinguish between species, es­ There is some disagreement as to which pecially the commercially important ones. North American species of Betula are com­ Wood anatomists have not previously dem­ mercially important. Brisbin and Sonderman onstrated methods for the reliable separation (1973) state that B. alleghaniensis, B. lenta, of the birches. Hall (1952) studied the wood and B. papyrifera are of primary significance of 33 species of Betula L., but offered no and B. nigra and B. populifolia are second­ means for their separation. A year later, Stark ary. Core et al. (1976, 1979) limit discussion (1953) published a key to the woods of six to B. alleghaniensis and B. nigra; where-as North American tree species of birch using Panshin and De Zeeuw (1970, 1980) and number of bars per perforation plate, vessel Panshin et al. (1964) include B. alleghan­ element and fibre lengths, heartwood colour, iensis, B. lenta, B. nigra, B. papyrifera, and and density as distinguishing features. This B. populifolia. We followed Brisbin and Son- Miller & Cahow - Wood identification of birch (Betula) 365 derman (1973) and selected B. alleghanien­ of Canada and west to Alaska. It extends sis, B. lenta, and B. papyrifera as the com­ south to Pennsylvania, Wisconsin, and into mercially important species. We included B. Idaho. It will occur as far north as Hudson’s nigra because Panshin and De Zeeuw (1980) Bay in the east and into the Northwest Terri­ and Core et al. (1979) also included this spe­ tories in the west (Little 1971). cies. Of the three other North American tree Following the classification of Rehder species, we excluded B. uber, B. occidenta­ (1949), the six North American tree species lis, and B. populifolia as commercially im­ of Betula alleghaniensis, B. lenta, and B. ni­ portant because of their limited range, size, gra are placed within subsection Costatae; and availability. whereas, B. papyrifera, B. occidentalis, andB. The properties and economic uses of B. populifolia are assigned to subsection Albae. alleghaniensis and B. lenta are similar (Pan- Hall (1952) reports that subsections Costatae shin & De Zeeuw 1980; USDA Forest Ser­ and Albae are very closely allied, with the vice 1987), and these species are often mixed Costatae most likely the more advanced of the in the trade. Although these two species are two. relatively difficult to work with hand tools, they can be easily machined. However, they Materials and Methods must be dried carefully to prevent checking For each of the four commercially im­ and warping. Betula alleghaniensis is among portant North American species, 10 or 11 the principal furniture woods of the United wood samples were selected from the States (Brisbin & Sonderman 1973). Betula MADw* wood collection. The samples ex­ papyrifera and B. nigra are lighter and rank amined are listed below by species. Speci­ considerably below B. alleghaniensis and B. mens are cited by species, collector, collec­ lenta in strength, stiffness, and hardness. tor’s number, and xylarium designation in They are relatively easy to dry and to work parentheses.Unvouchered specimens arecited with hand tools. Both species are used for only by xylarium designation in parentheses. pulp, fuel wood, and small crafts, and B. papyrifera is desirable for tooth picks, tongue Betula alleghaniensis: W. E. McQuilken, depressors, and coffee stirrers because of its BWCw 8666 (MADw 767); H.P. Brown, machineability, nearly white colour, and the BWCw 8002 (MADw 8723); F. Schmidt fact that it does not impart a colour, odour, or (MADw 10162); Watkins, USw 2230 taste. (MADw 19910); W.E. Johns (MADw Bentula alleghaniensis and B. lenta are mes­ 23625); H.H. Smith, F 72208 (MADw ic forest species native to the Appalachian 27319); H.H. Smith, F 72202 (MADw region. Betula alleghaniensis, however, is 27320); H.H. Smith, F 72210 (MADw more northern and extends into the Lake 27321); A. Okkonen 1 (MADw 40885); A. States and Canada. It occurs along the Appa­ Okkonen 2 (MADw 40886). - B. lenta: lachian Mountains as far as northern Geor­ Ostronder (MADw 944); (MADw 6179); A. gia, and to the north grows east to the Koehler (MADw 6466); (MADw 6698); Maritimes and west to Minnesota. Betula len­ H.S. Newins, BWCw 8046 (MADw 8765); ta grows as far south as Alabama and as far D.A. Kribs, BWCw 8172 (MADw 8912); north as Maine. It is bounded in the west by Northeastern For. Exp. Stat., (MADw the Ohio Valley and Lake Erie. Betula nigra 23594); US Bureau Forestry, F 01690 is a southern lowland species, and it occurs (MADw 27316); H.H. Smith, F 72198 in the southeastern United States. It grows to (MADw 27317); H.H. Smith, F 72200 the Atlantic Ocean in the east and to Mis­ (MADw 27318). - B. nigra: Allegheny souri, eastern Oklahoma, and eastern Texas For. Exp. Stat., BWCw 8602 (MADw 825); in the west. It extends north to New Jersey, Allegheny For. Exp. Stat., BWCw 8601 Ohio, and Iowa and occurs locally in Wis­ consin, New York, and New Hampshire. * MADw and SJRw are acronyms for the Betula papyrifera is a northern boreal spe­ wood collection at the Forest Products cies, growing east to the Maritime Provinces Laboratory, Madison, Wisconsin, U.S.A. 366 IAWA Bulletin n.s., Vol. 10 (4), 1989

(MADw 826); H. Nogle, BWCw 8723 The number of bars per scalariform per­ (MADw 1262); (MADw 1730); D.T. foration plate was counted from macerations. Dinsmore, BWCw 8073 (MADw 8783); We arbitrarily selected and assessed 25 intact B.H. Paul, BWCw 8097 (MADw 8804); B. vessel elements. One perforation plate per Spike, BWCw (MADw 8902); R.O. Marts element was examined. In addition to calcu­ (MADw 11294); J.W. Thieret 1289 (MADw lating an average of the 25, the 10 perforation 25417); H.H. Smith, F 72228 (MADw plates with the greatest number of bars were 27322); H.H. Smith, F 72230 (MADw selected and averaged. 27324). - B. papyrifera: (MADw 7022); Jahn e Martell, BWCw 8362 (MADw 9222); Results H.P. Brown, BWCw 8000 (MADw 8721); The description follows the IAWA Stan­ H.S. Newins, BWCw 8136 (MADw 8846); dard List of Characters (IAWA Committee E. Fritz, BWCw (8228) (MADw 8943); For. 1981; Miller 1981) and is based on the ex­ Exp. Stat., BWCw (MADw 8968); Haines e amination of B. alleghaniensis, B. lenta, B. Wangaard, BWCw 8505 (MADw 9430); nigra, and B. papyrifera. Characters not spe­ H.C. Sulerud (MADw 18391), W.E. Johns cifically mentioned are negative. Quantitative (MADw 23624); US Bureau Forestry, F characters are ranges of averages. Table 1 01686 (MADw 27327). lists the ranges of averaged for vessel For each of the other species, we survey­ diameters, vessels per square millimetre, ray ed available slides and selected additional height, vessel element length, and fibre specimens from the MADw and SJRw wood length. Table 2 lists the diagnostic characters collections, but in no case did we examine for identification. The Appendix is a dicho­ more than seven specimens per species. tomous key including the four described spe­ We cut sectioning blocks and samples for cies, the two commercially important Euro­ maceration from the outer portion of the sap­ pean species (B. pendula and B. pubescens), wood whenever possible to obtain mature and the two additional North American tree trunk wood. Standard microtechnique prac­ species (B. occidentalis and B. populifolia). tices were followed. General features - Heartwood off-white Observations and measurements follow the to brown, occasionally with slight reddish procedures described by Miller (1981) with tinge; not fluorescent under longwave (365 the exception of vessels per square milli­ nm) ultraviolet light. Pith flecks occasionally metre, the number of bars per perforation present in some specimens. Specific gravity plate, and the reporting of some quantitative (ovendry weight/green volume) 0.55 in B. values. For vessels per square millimetre, alleghaniensis, 0.60 in B. lenta, 0.48 in B. each vessel or opening is counted as a single papyrifera (USDA Forest Service 1987), and unit (Wheeler 1986). For vessel and fibre 0.49 in B. nigra (Lutz 1972). lengths, vessel diameters, and ray heights, Microscopicfeatures - Growth rings dis­ the range reported is the range of specimen tinct. Vessels diffuse, solitary, and in short averages. radial multiples of 2-3; 19-58/mm2, most-

Table 1. Range of averages for vessel diameter, vessels per square millimetre, ray height, vessel element length, and fibre length. Miller & Cahow - Wood identification of birch (Betula) 367

Table 2. Useful characteristicsfor the identification of the commercially important North Ameri­ can species of Betula.

ly round; average diameters 54-102 µm, bling septate (gum bars) may be conspic­ average vessel element length 503-1096µm uously present in some specimens; without (Table 1). Perforation plates exclusively sca­ conspicuous bordered pits; walls thin to lariform; on average 6-10 bars in B. nigra thick; average length 1094-1735µm (Table (Fig. 2), 10-20 bars in B. alleghaniensis and 1). Tracheids rarely present. Rays essentially B. lenta, and 15-26 bars in B. papyrifera homocellular but with occasional square or (Fig. 3; Table 2). Spiral thickenings absent; upright cells on the margins of the ray; on although intervascularpit apertures occasion­ average 170-404µm in height (Table 1); not ally coalescent sometimes giving the impres­ of two distinct sizes; not storied. Aggregate sion of helical sculpturing. Tyloses absent. rays absent. In B. alleghaniensis, B. lenta, Intervascular pitting alternate; mostly 4 µm, and B. nigra, rays mostly 3-4, occasionally occasionally 3 or 5 µm; vessel-ray pitting to 5, cells wide; in tangential longitudinal same as intervascular pitting; not vestured. section ray shape typically fusiform and ray Fibres nonseptate, although structuresresem­ cell shape typically circular (Figs. 4, 5; Table

Fig. 1. Scatter plot showing the average number of bars per perforation plate and the average ray width in cells. A = Betula alleghaniensis; L = B. lenta; N = B. nigra; P =B. papyrifera. 368 IAWA Bulletin n.s., Vol. 10 (4), 1989

Figs. 2 & 3. Radial sections showing number of bars per scalariform perforation plate; × 200. ­ 2: Betula nigra showing 11 bars/plate on the left and 8 bars/plate on the right. - 3: B. papyri­ fera showing 21 bars/plate on the left and 19 bars/plate on the right.

2). In B. papyrifera, rays 2-3,rarely to 4, Discussion cells wide; in TLS ray and ray cell shape The wood anatomy of all species of Betula laterally flattened (Figs. 6,7; Table 2). Apo­ is relatively homogeneous (Hall 1952). tracheal parenchyma marginal and diffuse, Panshin et al. (1964), Panshin and De Zeeuw scanty to frequent; paratracheal parenchyma (1970, 1980), and Core et al. (1976, 1979) absent; cell contents as viewed in unstained reported no diagnostic differences among the sections russet (reddish brown or yellowish commercially important North American brown) in B. alleghaniensis, B. lenta, and B. species. Stark (1953), however, studied the nigra (Fig. 8); colourless or without contents native tree species of the United States and in B. papyrifera (Fig. 9; Table 2). Prismatic published a key that we question. crystals absent. One druse in one sample of In his first couplet, Stark (1953) separates B. nigra (MADw 8902). B. nigra from the other species on the basis

Figs. 4-7. Tangential sections showing ray width, ray shape, and ray cell shape. - 4: Betula lenta showing rays 3-4cells wide and the typical fusiform ray shape; × 130. - 5: Ray from 4 at × 500 showing circular ray cell shape. - 6 B. papyrifera showing rays 2-3cells wide and the laterally flattened ray shape; × 130. - 7: Ray from 6 at × 500 showing laterally flattened ray cell shape. Miller & Cahow - Wood identification of birch (Betula) 369

Miller & Cahow - Wood identification of birch (Betula) 371 with wide rays and few bars. This clustering The shape of the rays and ray cells as is relatively tight, and the highest value we viewed on the tangential longitudinal section obtained for average number of bars is 10.5, is also diagnostic. In B. papyrifera, the rays which is relatively close to one specimen of and ray cells appear laterally flattened (Figs. B. alleghaniensis (10.8) and one specimen of 6, 7). and in B. alleghaniensis and B. lenta, B. lenta (10.8). Thus, Figure 1 indicates that the rays are typically fusiform and the ray the average number of bars is a diagnostic cells circular (Figs. 4, 5). Any of these four feature for separating B. nigra from the other features is not sufficiently diagnostic to ac­ three species. Betula lenta and B. alleghan­ curately separate B. papyrifera from B. alle­ iensis cluster in the middle of the range. ghaniensis and B. lenta, but all four used in Within their range, we observe that B. lenta combination are diagnostic. tends toward fewer bars and wider rays. The To determine whether or not the average trend, however, is not strong enough to per­ number of bars per perforation plate is more mit a positive identification. Betula papyrifera diagnostic than a biased average of 10 plates tends toward the bottom right comer, im­ with the highest number of bars, we exam­ plying a high number of bars with narrow ined the data and performed t-tests. We per­ rays. One specimen of B. papyrifera fills formed two-sided t-tests on each pair-wise within the B. alleghaniensis and B. lenta combination of species for the averages of cluster. This suggests that the number of bars all 25 plates and repeated this for the biased and ray width may not be sufficient to sepa­ averages of 10 plates with the highest number rate these taxa. of bars. In virtually every case, a higher Betula papyrifera is distinguished from B. resolution of the differences between species alleghaniensis and B. lenta using a combina­ was gained by using the average of all 25 tion of four characters (Table 2, Appendix). plates, indicating that the overall average of As previously noted, B. papyrifera most 25 plates is more diagnostic than the biased often has more bars per perforation plate than average of 10 plates with the highest number B. alleghaniensis and B. lenta (Table 2, Fig. of bars. 3), but this feature by itself is not sufficient We examined six specimens ofB. occiden­ to accurately separate the species. talis and seven specimens of B. populifolia, As shown in Figure 1 and Table 2, B. pa­ two additional tree species in North America pyrifera has slightly narrower rays than the (Little 1971), to determine if their wood anat­ other species. In B. papyrifera, the rays are omy is similar to the commercial species. mostly two to three cells wide at the widest Betula occidentalis is very similar to B. pa­ point (Figs. 6, 7). In our 10 specimens, we pyrifera and both are in the subsection Albae. found 3 specimens with rays four cells wide, In fact, B. occidentalis was once considered a but in each specimen fewer than three rays variety of B. papyrifera (Little 1971). It is not four cells wide were observed. This is based surprising then that we could not find on surveying the entire section (five measure­ any character that would separate the two. In ments in five fields) and selecting only the five of the six specimens, the rays and ray widest rays at the widest point. Rays of B. cells are laterally flattened; the rays 2-3 cells alleghaniensis and B. lenta are three to four wide; the axial parenchyma cell contents cells wide (Figs. 4, 5). In our 20 specimens, mostly colourless or absent; and the average we observed only 1 specimen where we number of bars per perforation plate ranges could not find at least one ray that was four from 14 to 21. One specimen collected with­ cells wide. out herbarium material in Mt. Shasta National The third differentiating character is axial Park in California, however, has all the char­ parenchyma cell contents. In B. papyrifera, acteristics of B. alleghaniensis and B. lenta, the axial parenchyma cells are without con­ i.e., copious coloured axial parenchyma cell tents or the contents are colourless (Fig. 9). contents, rays 3-4 cells wide, circular ray In contrast, B. alleghaniensis and B. lenta cells, and rays with a fusiform shape on the have russet (reddish brown or yellowish TLS. Perhaps this specimen is misidentified, brown) contents (Fig. 8). but B. occidentalis is the only tree species 372 IAWA Bulletin n.s., Vol. 10 (4), 1989 native to California. As unlikely as it may B. pubescens have rays 3-4 cells wide, and seem, this specimen may represent an ex­ the shape of the rays and ray cells are fusi­ treme in variability. Without additional speci­ form and circular, respectively. The number mens and closer scrutiny. we cannot explain of bars per perforation plate varies from 10to this apparent deviation from the norm. 21 and that overlaps the range of averages for Betula populifolis appears most like B. B. alleghaniensis, B. lenta, and B. nigra; how­ alleghaniensis and B. lenta, but sometimes it ever, the axial parenchyma cell contents are has characteristics of other species. The rays absent or colourless. Therefore, we can sepa­ are usually 3-4 (occasionally 2-3) cells rate B. pendula and B. pubescens from B. wide, but the shape of the rays and ray papyrifera on the basis of ray width and from cells vary from laterally flattened to fusiform B. alleghaniensis, B. lenta, and B. nigra on shaped rays and circular ray cells. The aver­ the basis of axial parenchyma cell contents age number of bars per perforation plate (Appendix). ranges from 15 to 24 which is suggestive of In conclusion, of the four commercially B. papyrifera; however, the axial parenchy­ important species of North America, B. nigra ma cell contents vary from russet to absent. and B. papyrifera can be identified from each Although this species was possibly of some other and from the two indistinguishable spe­ historical importance in northeastern United cies. B. alleghaniensis and B. lenta. Charac­ States, it is currently used primarily for fire­ ters used for these identifications include a wood and pulpwood and even that does not combination of ray width, ray and ray cell command much of the total market. Betula shape, average number of bars per perfo­ populifolia is a highly variable species; there­ ration plate, and the axial parenchyma cell fore, we group it with B. papyrifera, B. alle­ contents. In addition, the commercially im­ ghaniensis, B. lenta, and with the commer­ portant European species, B. pendula and B. cially important European species in our key pubescens, can be separated from the four (Appendix). North American species using the same com­ In Europe, the commercially important bination of characters. Two additional North birches are limited to B. pendula and B. pub­ American tree species, B. occidentalis and B. escens. According to Rehder (1949), these populifolia, are not distinct. Betula occiden­ species are included with B. papyrifera in talis cannot be separated from its close rela­ subsection Albae. Bhat and Karkkainen tive B. papyrifera, and B. populifolia is quite (1980) suggest a character that might separate variable and overlaps the anatomies of B. al­ these two woods. They argue that the aver­ leghaniensis, B. lenta, B. papyrifera, B. pen­ age number of bars per perforation plate is dula, and B. pubescens. indicative of species; B. pendula averaging less than 17.6 bars and those of B. pubes­ References cens averaging more than 17.6 bars. Al­ though our results support this trend, they did not reveal a discontinuity sufficient for identification. We measured the number of bars on 25 perforation plates for three speci­ mens of B. pubescens and four of B. pen­ dula that yielded the following range of aver­ ages: 10-16for B. pendula; 16-21 for B. pubescens. The breakpoint of 16 is suffi­ ciently different from 17.6 to raise doubts about separation of the European birches using this character. As a group, B. pendula and B. pubescens possess a combination of characters that seg­ regate them from the commercially important North American birches. Betula pendula and Miller & Cahow - Wood identification of birch (Betula) 373

Appendix: Dichotomous key to birches

The following dichotomous key includes four commercially important North American species (Betula alleghaniensis, B. lenta, B. nigra, and B. papyrifera), two commercially important European species (B. pendula and B.pubescens ), and two additional North American species (B. occidentalis and B. populifolia). The commercial North American species are in bold type, whereas the European and two additional North American species are in parentheses. It is im­ portant to emphasise the importance of using macerations to obtain accurate counts for bars per perforation plate and determining averages based on a sampling of 25 plates. The key is intended for use with reference specimens to aid in the interpretation of character states. 1. Rays 2-3cells wide; shape of rays and ray cells (TLS) laterally flattened; axial parenchyma cells without russet (reddish brown or yellowish brown) contents; average number of bars per perforation plates greater than 15 ...... Betula papyrifera (B.occidentalis) (B. populifolia) 1. Rays 3-4cells wide; shape of rays (TLS) fusiform and shape of ray cells generally circular; axial parenchyma cells with or without russet (reddish brown or yellowish brown) contents; average number of bars per perforation plates generally 6-22 ...... 2 2. Average number of bars per perforation plate 6-10 ...... Betula nigra 2. Average number of bars per perforation plate 10-20 ...... 3 3. Axial parenchyma cells with russet (reddish brown or yellowish brown) contents Betula alleghaniensis B. lenta (B.populifolia) 3. Axial parenchyma cells without russet (reddish brown or yellowish brown) contents (Betulapendula) (B.pubescens) (B.populifolia)