Sapindaceae gen. and sp. indet.

(Fig. 49)

EXAMINED MATERIAL. — Ethiopia • Lower Omo Valley, Shungura Formation, locality non available, sect. 22, Member G-29, 1.92 Ma; Frank Brown & Jean de Heinzelin leg; Twf 17966; Tervuren collection.

DESCRIPTION

The transverse and radial sections are fairly well preserved, while the longitudinal and tangential sections are poorly preserved. Diagnostic features are only distinct in limited areas of each section. The studied sample is a small stem with a central cylinder of parenchyma. The wood is diffuse-porous with distinct growth rings (Fig. 49A). Vessels are mostly solitary (84%) and few in radial multiples of 2 (11%), 3 (5%) (Fig. 49B). Mean tangential diameter of 80 µm, range 65-100 µm, 16-30 vessels per mm². Perforation plates are exclusively simple, inter-vessel pits not distinct, vessel-rays pits large and bordered (Fig. 49G). Axial parenchyma is scanty paratracheal and in marginal bands, initial/final parenchyma (one row), strands of parenchyma not seen in longitudinal section. Rays are uniseriate, few of them are biseriate on one row, composed of procumbent cells with one row of upright cells, 10-12 per mm2, mean ray height range is 240 µm (range 73-312 µm) (5-28 cells), mean ray width is 11 µm (range 10-12 µm) (Fig. 49C, D). Fibers are thick-walled and septate. Crystals are distinct in ray cells. Fibers are thick to mostly very thick walled, not clearly distinct in tangential view.

AFFINITIES AND COMPARISONS

The fossil is characterized by diffuse-porous wood with distinct growth rings. Vessels occur singly or in radial multiples of 2 or 3, featuring exclusively simple perforation plates. Axial parenchyma is scanty paratracheal, forming marginal bands and initial/final parenchyma. The general wood anatomy of the Shungura sample closely resembles that of the fossil genus Terminalioxylon (Mädel-Angeliewa & Müller-Stoll 1973), but it has small vessels with well-defined pits rather than vestured pits. This observation excludes affinities with Terminalia, as the presence of vestured pits is one of the diagnostic features of this genus (Van Vliet 1979). The Twf 17966 fossil specimen more likely corresponds to the structure of the family Sapindaceae (Klaassen 1999).

Micro-anatomical structure of Twf 17966 is near that of the morphotype Sapindoxylon Kräusel, described from the Cretaceous to the Plio-Pleistocene, in Africa as well as in Ethiopia. Sapindoxylon sp. Beauchamp (Cretaceous-Eocene) has large rays (Beauchamp et al. 1973). S. almelai Koeniguer (Mio-Pliocene Sahara) has more bigger vessels than Twf 17966 (Koeniguer 1967). S. haileselassiei Lemoigne (Oligo-Miocene from Ethiopia) has 1-3 seriate rays, vasicentric parenchyma and wider vessels in radial multiple up 15 vessels (Lemoigne1970). S. (Blighioxylon) koholense Koeniguer (Algeria) and S. sp. Koeniguer (Libya) have 1-3 seriate rays, large vessels and less vessels density than the Omo fossil and scanty paratracheal parenchyma (Koeniguer 1967, 1972). S. mbaense Koeniguer (Eocene, Senegal) has uniseriate rays and scanty paratracheal parenchyma (Koeniguer 1973a). S.stromeri Kräusel (Oligo-Miocene from Egypt) has vasicentric parenchyma, wider vessels in radial multiple up 15 vessels, septate fibers and prismatic crystals in rays and parenchyma, two features indistinct in Twf 17966. A fossil wood placed in Sapindoxylon, related to the type specimen S. stromeri has been described from the site Fwjj 14, dated 1.56 to 1.36 Ma at Ileret, East Turkana, Kenya. It differs from Twf 17966 by having bigger vessels, smaller radial multiples, exclusively uniseriate rays and parenchyma that is either rare or sparse paratracheal (Bamford 2017). In conclusion, Twf 17666 is attributed to the Sapindaceae without any generic precision (Table 2; Fig. 50).