Klaassenoxylon, 2023

Wheeler, Elisabeth A., Manchester, Steven R. & Baas, Pieter, 2023, A late Eocene wood assemblage from the Crooked River Basin, Oregon, USA, PaleoBios 40 (14), pp. 1-55 : 35-38

KLAASSENOXYLON WILKINSONII SP. NOV.

FIG. 17A–K

Diagnosis— Growth rings distinct to indistinct. Wood diffuse-porous. Narrow vessels solitary and in short radial multiples. Perforation plates exclusively solitary; intervessel pits alternate, minute to small; vessel-ray parenchyma pits similar to intervessel pits. Fibers with simple pits, septate and non-septate. Axial parenchyma rare, scanty paratracheal. Rays uniseriate (rarely biseriate), heterocellular, less than 1 mm tall. Long chains of crystals in fibers,

Holotype — UF 278-84877 , estimated maximum diameter 12 cm.

Paratypes — UF 278-84882 , 84897 , estimated maximum diameters 2.2 cm (with pith), 4 cm .

Etymology— Named for Hazel Wilkinson (1932- 2019), a devoted plant anatomist with a long career at the Jodrell Laboratory, Kew, U.K.

Occurrence— Dietz Hill ( UF 278).

Description— Growth rings present, marked by radially narrow fibers ( Fig. 17A, B ).

Wood diffuse-porous; vessels solitary 29% in UF 278- 84877, 64% in UF 278-84897 and in radial multiples of 2–3(-4), solitary vessels round to oval in outline ( Fig. 17A, B ); mean tangential diameters 74 (16) µm in UF 278-84877, 67 (12) µm in UF 278-83897, total range of both specimens 45–106 µm; 18–23–33 vessels/ mm 2. Perforation plates exclusively simple ( Fig. 17C, F ); intervessel pits alternate, minute to small ( Fig. 17C ); vessel-ray parenchyma pits similar to intervessel pits ( Fig. 17D ). Average vessel element length 350 ( SD =71) µm, range 249–452 µm.

Axial parenchyma rare, scanty paratracheal, in 3–4-celled strands as viewed in radial section.

Fibers thin- to thick-walled, pitting not observed, septate and non-septate ( Fig. 17F, G ).

Rays uniseriate ( Fig. 17F, G ), with intermixed weakly procumbent to square cells to weakly upright cells; in others a tendency for square to upright cells to be restricted to ray margins; average ray height 346 ( SD =154) µm in UF.84877; 284 (127) µm in UF 278-84897; 10–14 rays per mm.

Prismatic crystals in chambered fibrous elements.

Description of pith— Pith complex, surrounded by a sheath of small cells with dark contents and clusters and strands of fiber-sclereids and sclereids ( Fig. 17I ). Pith proper composed of two types of parenchyma cells: 1) large “empty” rectangular cells that are in radial rows towards the periphery and clusters of isodiametric cells in the center, and 2) small cells with dark contents (tannins?) forming a three-dimensional network throughout the pith and partly segmenting it by concave arches as seen in cross section ( Fig. 17G ). There are isodiametric empty cells and dark cells resembling irregular “rosettes” as described by Poole and Wilkinson (1992) for Sapindoxylon guioaoides Poole and Wilkinson (1992), Guioa Cav. (1797) , and Alectryon Gaertn. (1788) . However, UF 278-84882’s pith differs significantly in having a distinct sheath of small-cells including some sclereids bordering the protoxylem. Prismatic crystals present in peripheral pith parenchyma ( Fig. 17J, K ).

Comparison to extant woods— We initially searched the InsideWood database for diffuse-porous wood (5p), vessels solitary and in short multiples and not arranged in an obvious pattern (6–11a), simple perforation plates (13r), minute to small alternate intervessel pitting (22p, 26a, 27a), vessel-ray parenchyma pits similar to intervessel pits (30p), vessel tangential diameter averaging less than 200 µm (43a), vessel frequency between 5–40 per mm 2 (46a, 49a, 50a), fibers with simple pits (61p), axial parenchyma not obvious (77a, 79–86a), rays exclusively uniseriate (96p), storied structure absent (118a, 120a), and prismatic crystals present in fibers (143p). This first search returned two Lythraceae J. St. -Hil. (1805) records and 37 Sapindaceae .

Search 2 included presence of both septate and non-septate fibers and yielded one Lythraceae and 32 Sapindaceae .

Search 3 added presence of rays with procumbent, square, and upright cells intermixed throughout the ray (109p) and absence of dimorphic fibers (67a) this nar- rowed the possibilities to seven sapindaceous genera, one of which, Lecaniodiscus Planch. ex Benth. (1849) has crystals in ray cells as a constant and so does not match the Dietz Hill ( UF 278) woods. Consequently, we consider that these woods are allied with the Sapindaceae .

Klaassen (1999: p. 174–175) surveyed the wood anatomy of the Sapindaceae and recognized four wood anatomical types. These Dietz Hill ( UF 278) woods fit his Type IVa, which includes 26 genera that have intervessel pits of 3–6 µm, septate fibers, scanty paratracheal parenchyma, heterocellular uniseriate rays, and crystals in fibrous elements. Cnesmocarpon Adema (1993) New Guinea, Australia), Delavaya Franch. (1886) ( China), Gereaua Buerki and Callm. (2010) (syn. Haplocoelum Radlk., 1878 ; Africa), Macphersonia Blume (1848) (Tropical E. Africa, Madagascar), Pseudopteris Baill. (1874) ( Madagascar), Ungnadia Endl. (1833[1835]; Mexico, sw USA). Klaassen’s synoptical key with twenty-six features is slightly awkward to use. We excluded genera that have as constant features ones not observed in these fossils (vessels in a radial pattern, mean intervessel pit diameter>7 µm, mean vessel diameter>120 µm, fiber dimorphism present, apotracheal parenchyma abundant, complete sheaths of vasicentric parenchyma, aliform to conflu- ent parenchyma, banded non-marginal parenchyma, idioblastic axial cells, bi- and tri-seriate rays common, mean ray height> 1 mm, rays all homocellular, rays of two distinct sizes, crystals in ray cells). In addition to the six genera listed above as sharing features with UF 278-84887, Klaassen’s key indicates the following genera also have similarity: Arytera Blume (1847) , Beguea Capuron (1969) , Cupania L. (1753), Cupaniopsis Radlk. (1879/1880), Diploglottis Hook.f. (Bentham and Hooker, 1862) , Neotina Capuron (1969) , Thouinia Poit. (1804) , and Toechima Radlk. (1877[1879]) of Type IVa; Alectryon Gaertn. (1788) , Matayba Aubl. (1775) , Schleichera Willd. (1805) of Type IVb; and Handeliodendron Rehder (1935) , which did not fit in any of Klaassen’s wood type groups.

Buerki et al.’s (2021) recent analysis of Sapindaceae recognized four subfamilies and twenty tribes. All six genera from InsideWood Search 3 belong to subfamily Sapindoideae Burnett (1835) , but to three different tribes: Ungnadieae Buerki and Callm. (2021) ( Delavaya , Ungnadia ), Stadmanieae Buerki and Callm. (2021) ( Gereaua Buerki and Callm. (2010) , Macphersonia , Pseudopteris ), and Cupanieae Blume ( Cnesmocarpon ).

It is tempting to suggest, based on present day geographic distribution as well as shared anatomical features, that these Dietz Hill sapindaceous woods might be related to tribe Ungnadieae . However, as pointed out by Klaassen (1999), wood anatomical differences between the genera of groups IVa and IVb are small and could easily break down upon examination of more species per genus and more specimens per species. The main conclusion of the searches reported above and assorted alternative searches we did consistently yielded multiple Sapindaceae . We consider the similarity in the complex pith anatomy with some modern and fossil Sapindaceae as strong support for the family assignment.

Comparison to fossil woods— The diagnosis of Sapindoxylon as adapted by Poole and Wilkinson (1992) from a translation of Kraüsel’s work (1922) follows: “ Wood: Diffuse porous. Growth rin gs: lacking. Vessels: 2–8 per mm 2, mostly solitary or paired, seldom in threes, when abutting tangentially flattened, otherwise circular or el- liptical. Perforation plates simple with mainly horizontal or slightly oblique end walls. Fibres: Libriform, forming the groundmass in regularly arranged rows; occasionally septate. Wood parenchyma: Surrounding the vessels in one or two layers [vasicentric], sometimes also scattered. Rays: Uniseriate, occasionally biseriate, 3–30 cells high, consisting of one cell type; the marginal cells occasionally contain a crystal.”

It is unfortunate that the vessel frequency was so narrowly defined. Some Sapindoxylon species do not conform to the diagnosis in that feature, e.g., Sapindoxylon guioaoides (average of 40 vessels mm 2), S. elattostachyoides Grambast-Fessard (1966) , (20–40 vessels per mm 2), S. koelreuteroides Poole and Wilkinson (1992) (average of 43 vessels mm 2). These Dietz Hill ( UF 278) samples have double the number of vessels per mm 2 of Sapindoxylon Kraüsel. They also differ in 1) not having vasicentric axial parenchyma, but just scanty paratracheal; 2) heterocellular rays, rather than homocellular; 3) crystals in fibers/ chambered axial parenchyma, rather than in marginal ray parenchyma cells. These three features were used in Klaassen’s (1999) synoptic key and so were considered important in recognizing genera. Sapindoxylon klaassenii Wheeler and Manchester (2002) from the middle Eocene Clarno Nut Beds differs from the Kraüsel’s (1922) diagnosis in having rare axial parenchyma and heterocellular rays.

Consequently, we choose not to assign UF 278-84887 to Sapindoxylon Kraüsel and propose the genus Klaassenoxylon to accommodate these Eocene sapindaceous woods from Oregon. We also propose the new combination:

Figures

Figure 17. Caption on pg. 37.