A late Eocene wood assemblage from the Crooked River Basin, Oregon, USA Author Wheeler, Elisabeth A. Author Manchester, Steven R. Author Baas, Pieter text PaleoBios 2023 2023-11-01 40 14 1 55 http://dx.doi.org/10.5070/p9401462457 journal article 304161 10.5070/P9401462457 55f81c30-f18f-44ef-a480-dc7531b3b818 0031-0298 10913330 HAMAMELIDOXYLON LIGNIER, 1907 CF. HAMAMELIDOXYLON SP. FIG. 21A–H Description— Wood diffuse-porous, narrow growth increments, boundaries of 1–3 radially narrow fibers ( Fig. 21A, B ). Vessels predominantly solitary, overlapping end walls of vessel elements appearing as tangential or oblique pairs; vessels angular in outline ( Fig. 21A, B ); very narrow to narrow, average tangential diameter 43 (9) µm; very numerous,> 100 mm 2 . Perforation plates exclusively scalariform with more than 20 bars ( Fig. 21C, D ), 34–42 bars counted. Intervessel pits scalariform to opposite. ( Fig. 21E ) Vessel-ray parenchyma pits horizontally elongate, half-bordered or with slightly reduced borders ( Fig. 21F ). Widely spaced tyloses common. Figure 20 . Caption on pg. 45. Figure 20. Malpighiaceae ?. UF 278-84892. A‒C . Distinct growth ring boundaries, wood diffuse-porous, vessels predominantly in radial multiples, TS. D . Simple perforation plates (PP), ray cells predominantly procumbent. RLS. E . Scalariform perforation plate with 5 bars, RLS. F. Perforation plate with partial bars, RLS. G. Crowded alternate intervessel pitting. RLS. H. Vessel-ray parenchyma pits. RLS? I. Tyloses in vessels, ray cells predominantly procumbent. RLS. J, K . Rays uniseriate and partially biseriate, axial parenchyma strands. TLS. L . Short vessel element, simple perforation plate (PP), rays partially biseriate. TLS. Scale bars=500 µm in A; 200 µm in B; 100 µm in C, I, J; 50 µm in D, K, L; 20 µm in E‒H. Figure 21. Incertae Sedis. cf. Hamamelidoxylon sp. , UF 278-84876. A, B . Wood diffuse-porous, vessels solitary, angular in outline, TS. C, D . Scalariform perforation plates, RLS. E . Transitional intervessel pitting,,scalariform to opposite, TLS. F. Vessel-ray parenchyma (VRP) pits horizontally elongate, RLS. G . Ray cellular composition, intermixed square, upright, and barely procumbent cells. H. Rays exclusively uniseriate, T next to beginning of tyloses formation, TLS. Scale bars=200 µm in A; 100 µm B, H; 50 µm in C; D, G; 20 µm in E, F. Fibers non-septate, thin-walled, bordered pits on radial walls. Axial parenchyma diffuse and scanty paratracheal, in strands of 5 or more cells ( Fig. 21H ). Rays uniseriate ( Fig. 20H ), heterocellular, composed of intermixed procumbent, square, and upright cells ( Fig. 21G ); 9–14 rays/mm. Crystals and oil/mucilage cells not observed. Specimen— UF 278-84876, estimated maximum diameter 4–5 cm . Occurrence— Dietz Hill (UF 278). Comparison to extant woods— We did multiple searches of the IW database for modern woods; all searches included diffuse-porous wood (5p), predominantly solitary vessels angular in outline (9p, 12p), exclusively scalariform perforation plates with more than 10 bars (13a, 14p, 15a), scalariform intervessel pits (20p), narrow numerous vessels (42a, 43a, 46a, 47a, 48a), non-septate fibers with bordered pits (62p, 66p) axial parenchyma not obvious (77a, 79–89a), exclusively uniseriate heterocellular rays (96p 104a, 105a). These searches returned Hamamelidaceae , usually Chunia H.T. Chang (1948) , Exbucklandia R.W. Br. (1946) , Hamamelis Gronov. ex L. (1753), and Theaceae Mirb. (1816) , usually Franklinia W. Bartram ex Marshall (1785) , Gordonia Ellis (1770) , Schima Reinw. ex Blume (1823) . Although in different orders, Saxifragales Bertchtold and J. Presl. (1820) and Ericales Bertchtold and J. Presl. (1820) , respectively, these families have some genera with similar combinations of wood anatomical features, making it difficult, often impossible, to determine the familial affinities of small, isolated pieces of wood. The anatomy of the Theaceae has been reviewed by Liang and Baas (1990 , 1991 ); the anatomy of the Hamamelidaceae has been reviewed by Tang (1943) and Wheeler et al. (2010) . Descriptions of species of both families appear in wood anatomical atlases (e.g., Pearson and Brown 1932 , Metcalfe and Chalk 1950 , Suzuki, et al. 1991, Sosef et al. 1998 , Itoh et al. 2022 , and the online FFPRI Wood Database, https://db.ffpri.go.jp/WoodDB/index-E.html). Comparison to fossil woods— A search of the IW fossil wood database using the same features given above returned species assigned to Corylopsites Mathiesen (1932) and Hamamelidoxylon Lignier (1907) . The description of Corylopsites groenlandicus Mathiesen (1932) lacks details on vessel-ray parenchyma pitting, vessel element frequency; it differs in having crystals in both ray and axial parenchyma cells. Wheeler and Manchester (2021 , Table 2 ) compared Hamamelidoxylon species with reasonably complete descriptions. Of those species, the most similar to this Dietz Hill (UF 278) wood is H. uniseriatum Wheeler and Manchester (2002) from the middle Eocene Clarno Nut Beds (but this species usually has rays with procumbent body cells, not intermixed procumbent, square, and upright cells. One of the results, cf. Hamamelidoxylon sp. Terada et al. (2011) , was considered “Family incertae sedis” . The authors noted that this wood type had features seen in the Hamamelidaceae and Theaceae , as well as Illiciaceae A.C. Sm. (1947) . Illiciaceae wood has narrow rays, but not exclusively uniseriate rays and so is not similar to the Dietz Hill (UF 278) fossil. We follow Terada et al.’s example and note that while UF 278-84876 has a combination of features that are consistent with Hamamelidoxylon , we cannot assign it to family. We suggest that some other Hamamelidoxylon species need reassessment to determine whether or not they have combinations of features unique to the Hamamelidaceae .