IAWA Journal, Vol. 18 (4), 1997: 331-368 WOOD ANATOMY OF THE HIPPOCRATEOIDEAE (CELASTRACEAE) by Alberta M.W. Mennega Department of Plant Ecology & Evolutionary Biology, Herbarium Division, University of Utrecht, Heidelberglaan 2, P. O. Box 80.102,3508 TC Utrecht, The Netherlands SUMMARY In this paper the wood anatomy of the subfarnily Hippocrateoideae of the Celastraceae is treated. Halle's division (1986, 1990) of the subfarnily into four tribes, chiefly based on material of tropical Africa: viz. Salacieae, Campylostemoneae, Helictonemeae and Hippocrateae is followed. In a recent issue of the Flora of the Guianas the Hippocrateaceae - there treated as aseparate farnily - were divided into Hippocrateoideae and Salacioideae. This bipartition was reflected in the wood structure of the genera studied (Mennega 1994). Here the wood structure of all gen­ era worldwide (24), except the Asian genus Arnicratea, is described. It appeared that again a subdivision into two distinct anatomical groups could be made, with the three last tribes mentioned above showing the same characteristic structure as found before in New World Hippocra­ teae/Hippocrateoideae. The most important features of this group are the presence of very wide and very high rays, in a number of genera with unlignified ray cells at the growth ring border, the absence of included phloem tissue, and in many species an intruding bark resulting in an in­ dented wood pattern in stern cross sections or even an intricate pattern of deep furrows. The Salacieael Salacioideae on the other hand are char­ acterized by narrow, not exceptionally high rays, absence of unlignified ray cells, the occurrence of septate fibres in a parenchyma-like dis­ tribution, and often by the presence of included phloem tissue, either as isolated strands or more often as conspicuous concentric bands, or as ir­ regular bands with radial connections. Features present in all genera are: vessels with simple perforation plates, preponderance of solitary vessels, wide and narrow vessels distributed at random, alternate pitting; fibre­ tracheids, and libriform nonseptate and septate fibres present; axial pa­ renchyma scanty paratracheal or as rare isolated strands; rays hetero­ geneous, the cell types irregularly distributed, rhombic crystals numerous, often in characteristic radial distribution. Campylostemon, considered in the past by some taxonornists as belonging in Celastraceae or as inter­ mediate between Hippocrateaceae and Celastraceae, closely resembles Hippocrateae in its wood anatomy. And it is especially this group that by its characteristic features -like the wide rays - is more different from Celastraceae in general than Salacieae, which have several features in common with genera of Celastraceae. Downloaded from Brill.com10/07/2021 06:13:28AM via free access 332 IAWA Journal, Vol. 18 (4),1997 Key words: Celastraceae, Hippocrateoideae, Salacieae, Campylo­ stemoneae, Helictonemeae, Hippocrateae, systematie wood anatomy, in­ cluded phloem, non-lignified ray tissue. INTRODUCTION The Hippocrateaceae, at present by most taxonomists considered as a subfamily of the Celastraceae, are confined to the tropics and subtropics. The plants are for the greater part woody lianas, occasionally shrubs or trees. A study ofthe New World genera (Mennega 1972, 1994) revealed the occurrence of two fundamentally different types of wood structure, running parallel with the taxo­ nomie division of the Hippocrateaceae family into two groups: Hippocrateae and Salacieae. In these papers the Hippocrateaceae were still considered aseparate family of the order Celastrales. It seemed worthwhile to study the wood structure worldwide to see if this bipartition is also found in genera not represented in the New World. Of particular interest was the structure of the African genera Bequaertia, Campylostemon and Tristemonanthes. Loesener, in his treatment of Hippocrateaceae (1942), considered Campylostemon as a genus of the Celastraceae, Bequaertia as a species of Hippocratea, and Tristemonanthes as a genus of the Hippocrateaceae. In a monograph of the West African Hippocrateaceae (1962) and in recent treatments in the Floras of Gabon and Cameroon by Halle (1986, 1990) these three genera were accommodated in the tribe Campylostemoneae of the Celastraceae. In spite of requests to several institutional wood collections no wood could be ob­ tained of Halles (1984) genus Arnicratea comprising fOUf Old World species. This genus is related to Reissantia (Hou 1963), but in Halles opinion sufficiently distinct to be considered as a separate taxon. Generic delimitations vary considerably in the subfamily Hippocrateoideae accord­ ing to the taxonomic opinions of the authors. Some prefer to admit only few genera, chiefly Hippocratea and Salacia, others are in favour of splitting these large genera into several smaller taxa. For my treatment of the New World material I followed A. C. Smith (1940), who ac­ cepted several genera next to Hippocratea and Salacia. For the African sampies the identification and papers by N. Halle provided the starting-point. He is not in favour of splitting Salacia, but on the other hand recognized several genera besides Hippocratea. Finally, for the Asian material Hou's (1964) treatment in Flora Malesiana was used. He considered the genus Salacicratea Loes. as conspecific with Salacia. From the present wood anatomieal research it appears that the tribes recognized by Halle in the Flore du Cameroun, listed in Table 1, can be divided into two groups: group 1 the Salacieae; group 2 comprising Campylostemoneae, Helictonemeae and Hippocrateae. Taxa of group 2 showing the characteristic wide rays of the Hippocrateae as described in the treatment of the Guianan material (Mennega 1994) will be indi­ cated in this paper as Hippocrateae. Downloaded from Brill.com10/07/2021 06:13:28AM via free access Mennega - Wood anatomy of the Hippocrateoideae 333 Table 1. Genera with number of species recognized, based on Halle (1990) and Görts & Mennega (1994); between brackets the number of species studied. Tribe Salacieae Campylostemoneae Helictonemeae Hippocrateae Cheiloclinium 11 (5) Bequaertia 1 (1) Helictonema 1 (1) Anthodon 2 (1) Peritassa 13 (5) Campylostemon 8? (1) Apodostigma 1 (1) Salacia c. 200 (31) Tristemonanthes 2 (1) Arnicratea 3 (0) Salacighia 2 (1) Cuervea 5 (1) Thyrsosalacia 4 (2) Elachyptera 7 (2) Tontelea 31 (6) Hippocratea 3 (3) Hylenaea 3 (2) Loeseneriella 16 (6) Prionostemma 5 (l) Pristimera 24 (8) Reissantia 6 (1) Semialarium 2 (1) Simicratea 1 (l) Simirestis 6 (1) MATERIAL AND METHODS For the present study a good number of sampies from Africa was available, main­ ly collected by F.O. Breteler and by A.lM. Leeuwenberg in Gabon and Cameroon. These collections were cited in Halles treatments for the Flore du Gabon and Flore du Cameroun (1986, 1990), where, for a number of specimens, a drawing of a cross section of the twig or stern at natural size was given, often showing characteristic pat­ tems. Material available from Asia and Australia was rather scarce, and in the case of Australia restricted to sections forwarded by the Wood Technology and Forest Re­ search Division at Beecroft, New South Wales. All sampies .are backed by herbarium vouchers. African herbarium specimens were identified by Halle, Neotropical material by myself. Wood sampies of the genera listed in Table 1 were studied. Sections were prepared according to standard methods. Samples or slides received from other institutes are indicated by Stem's (1988) abbreviations in his Index Xylariorum 3. Descriptions mainly followed the recommendations of the IAWA (1989). Features present in all taxa of the subfamily, e. g. simple vessel perforations and vessel / ray pit­ ting similar to the intervessel pitting, were omitted in the generic descriptions. Length of elements was measured in macerations, means were based on 25 measurements of each element. For vessel elements tails were included. In the short general descriptions of wood and bark only occasionally values could be given of the wood density. Downloaded from Brill.com10/07/2021 06:13:28AM via free access 334 IAWA Journal, Val. 18 (4), 1997 Generic treatments follow tribai succession as listed in Table 1, and within the tribes arrangement according to alphabeticalorder. Gregory's (1994) Bibliography was helpful in complementing and checking litera­ ture references. SURVEY OF WOOD ANATOMICAL CHARACTERS Growth rings Usually present but faint and hardly visible to the naked eye. Marked by presence of some rows of radially flattened fibres at the end of a growth period (Fig. 18,20). In several genera of the Hippocrateae (Table 2), however, the zone of flattened fibres extending in the wide rays as a V-shaped layer of unlignified cells, resulted in an unu­ sual pattern as seen in cross sections (Fig. 21,24,42). Vessels Generally for 90-95% solitary, the remainder in radial multiples of2-3; number of multiples slightly more numerous in Campylostemon and in some species of Salacia. Diffuse. Vessel frequency 4 -80 per sq. mm. The characteristic occurrence of wide and narrow vessels, Carlquist's (1985) vessel dimorphism (Fig. 1,24) is more pronounced in the lianas than in the shrubs or trees. Narrow vessels 20-50 Iffil wide, the wide ones ranging from 100-350 Iffil. Average vessel element length between 500-900 11m, mostly 600-800 Iffil. Perforations simple. Intervessel pits alternate, crowded, borders round to oval, the slits included or reaching the border or coalescent, suggesting a striated vessel