FLORA MALESIANA SERIES I—- SPERMATOPHYTA Flowering Plants Volume 11, part 2 Rosaceae, Amaryllidaceae, Alliaceae Coriariaceae, Pentastemonaceae, Stemonaceae LIBRARY THE NEW Yes BOTANICAL CARDEN BRONX Now ei 10458 FLORA MALESIANA SERIES I — SPERMATOPHYTA Volume 11 — part 2 — 1993 Rosaceae (C. Kalkman — pp. 227-351) Amaryllidaceae (D.J.L. Geerinck — pp. 353-373) Alliaceae (J.R.M. Buijsen — pp. 375-384) Coriariaceae (B.E.E. Duyfjes — pp. 385-391) Pentastemonaceae (B.E.E. Duyfjes — pp. 393-398) Stemonaceae (B.E.E. Duyfjes — pp. 399-409) CIP-GEGEVENS KONINKLIJKE BIBLIOTHEEK, DEN HAAG Flora Flora Malesiana. Series I, Spermatophyta : Flowering plants. - Leiden : Rijksherbarium / Hortus Botanicus, Leiden University Vol. 11, pt. 2: Rosaceae / C. Kalkman. Amaryllidaceae / D.J.L. Geerinck. Alliaceae /J.R.M. Buijsen. Coriariaceae / B.E.E. Duyfjes. Pentastemonaceae / B.E.E. Duyfijes. Stemonaceae / B.E.E. Duyfijes. - Il. Comp. and publ. under the auspices of Foundation Flora Malesiana. - Met index, lit. opg. ISBN 90-71236-19-6 Trefw.: flora ; Zuidoost-Azié. All rights reserved © 1993 Foundation Flora Malesiana No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without written permission from the copyright owner. Flora Malesiana ser. I, Vol. 11 (2) (1993) 227—351 ROSACEAE (C. Kalkman, Leiden, The Netherlands) Rosaceae Juss., Gen. Pl. (1789) 196, nom. cons.; Hutch., Gen. Flow. Pl. 1 (1964) 174-216; Vidal, Fl. Camb., Laos & Vietnam 6 (1968) 1—210 (excl. Rubus); Nguyen Van Thuan, ibid. 7 (1968) 1-83 (Rubus); Vidal, Fl. Thailand 2 (1970) 31-74 (Rubus by Nguyen Van Thuan); Tirvengadum, Fl. Ceylon 3 (1981) 328-378. — Type genus: Rosa L. Woody or herbaceous plants. Leaves usually spirally arranged, sometimes distichous, rarely opposite (not in Malesia), simple or compound. S$ tipules on the twig or on the base of the petiole, free or adnate to petiole, rarely absent. Inflorescences various. Flowers usu- ally bisexual and actinomorphic. Hypanthium (‘calyx tube’ of many authors) usually very distinct, from saucer-shaped to tubular or campanulate, the sepals, petals, and stamens inserted on its rim, its inside usually lined by a nectariferous disk. Sepals usually 5, free, in some tribes an epicalyx also present. Petals usually 5, free, from large and showy to small and not or hardly distinct from sepals, in some genera/ species absent. Stamens usually numerous, but sometimes the number distinctly related to the number of perianth leaves, filaments free, anthers bilocular, dehiscing longitudinally. Pistil(s) 1 to many, free or variously connate with each other and/or with the hypanthium, ovary(ies) superior to inferior, style(s) present, ovule(s) 1 to several (often 2) per locule, anatropous, ascending or pendulous. Fruits various, fleshy or dry, dehiscent or not. Seed(s) 1 to several, with- out or with scanty endosperm, cotyledons fleshy or flat. Distribution — A large family with worldwide distribution, including more than 3000 species in c. 100 genera. Almost all genera which are represented in Malesia have their ac- tual centre of distribution in temperate to subtropical regions on the Northern Hemisphere. Some of those are large or medium large genera with only one or two species in Malesia (Rosa, Alchemilla, Eriobotrya), others have a more or less distinct sub-centre in the Male- sian region (Prunus, Rubus, Potentilla). Exceptional is Acaena, a genus with a Southern Hemisphere distribution, of which one species also occurs in New Guinea. Kalkman [Bot. J. Linn. Soc. 98 (1988) 37—59] postulated a Southern (Gondwanan) origin for the family and migration via three routes. Malesia in this view was reached mainly from the Asian continent (Laurasia), which in turn was reached by way of South, Central, and North America and via Beringia. Partly maybe the continent was also reached directly from Gondwana by transport on the Indian ‘raft’ (Alchemilla?). A third route was via Australia (Acaena). Most authors, however, favour a Laurasian origin for the family. Habitat — The majority of Malesian Rosaceae belongs to the mountain flora and occurs only above 1000 a 1500 m altitude, in montane forest types, thickets or (sub)alpine grass- lands. Only in the genus Rubus (a dozen species) and in Prunus (more than 20 species) an appreciable proportion of the species are (also) found in the lowlands. (227) 228 Flora Malesiana ser.I, Vol. 11 (2) (1993) Ecology — As is true for almost all Malesian higher plant families, no autecological research has been carried out for members of Rosaceae. From the habitats where the spe- cies have been collected, some superficial conclusions may be drawn about preferences or tolerances for light, temperature and soil conditions and wherever possible, the paragraphs on Habitat and Ecology contain this kind of information. Pollination is undoubtedly normally by (unspecified) insects, as in the European rela- tives. Apart from the formation of a good quantity of pollen and the secretion of nectar by the disc, there are no specializations in the flowers related to pollination by specific kinds of insects. Only for Acaena wind-pollination might be inferred, but experimental or obser- vational evidence is lacking in literature or on labels. For dispersal most Rosaceae rely heavily on animals. Exceptions are found in genera with multi-seeded follicles with dry seeds (in Malesia only Neillia) where dispersal is by ballistochory. The same is true for most of the Potentilla species that have the dry achenes in the cups formed by the hypanthium, sepals and epicalyx. Some genera have dry achenes, imbedded in or surrounded by a fleshy spurious fruit (Rosa and also the not indigenous Fragaria and Potentilla indica). In these cases the hypanthium, resp. the torus functions as the attractant for endozoochory by snails, birds, or other animals. Many Rosaceae of dif- ferent tribes have gone the way to fruits with a fleshy or juicy layer in their walls (drupes, either single or as collective, or pomes) and obviously these are also endozoochorous. Epizoochory is only exercised by Acaena and Agrimonia which possess spines on their hypanthium in which the fruit is included. Taxonomy and Phylogeny — In modern systems the Neuradaceae and the Chryso- balanaceae [for the latter, see Flora Malesiana I, 10 (1989) 635—678] are mostly not in- cluded in Rosaceae, as in earlier classifications, but recognized as families in their own right. In the family Rosaceae as implied in the previous paragraph usually four subfamilies are distinguished: Spiraeoideae, Rosoideae, Maloideae (Pomoideae ), and Prunoideae (Amygdaloideae). The last-mentioned two groups are undoubtedly two end-branches in the phylogenetic tree, well recognizable, distinct, and natural (holophyletic) taxa. This cannot be said for the two other subfamilies. The group Spiraeoideae contains the genera with dry, dehiscent fruits. Dehiscent follicles is a plesiomorphic (primitive) character in the family and the genera possessing this character should better be included in a taxon with the genera that have been derived from them. Considering the likeness of the flowers of some Spiraeoideae and those of some Maloideae like Cotoneaster and Pyracantha, | would be inclined to enlarge the subfamily Maloideae with at least part of the Spiraeoid genera. The Rosoideae are quite heterogeneous and they have probably to be united with other, ‘Spiraeoid’, genera to form another holophyletic branch, in which maybe some subdivision is possible. A phylogenetic analysis, only considering morphological characters [Kalkman, Bot. J. Linn. Soc. 98 (1988) 37—59] was not successful and should be repeated with an augmented set of characters, also anatomical and chemical ones. Awaiting this, the recognition of the four classical subfamilies is hardly justified. Kalkman — Rosaceae 229 The next lower level of classification was used by Hutchinson, l.c., who divided the family in some twenty tribes. Some of these are heterogeneous, some others contain only one or two ‘difficult’ genera that have not yet found a good place with their nearest rela- tives in the phylogenetical sense. The firm core of a tribal classification consists of the following tribes: Spiraeeae (see p. 244) Neillieae (see p. 245) Gillenieae (maybe to be split into two tribes) (not in Malesia) Rubeae (see p. 247) Potentilleae (see p. 285) Dryadeae (probably to be divided into two tribes) (not in Malesia) Poterieae (see p. 297) Alchemilleae (usually included in Potentilleae or Poterieae) (see p. 301) Roseae (see p. 303) Maleae (Pomeae) (see p. 306) Pruneae (including Osmaronia?) (see p. 319) VZEMDADDDAAYYY S = tribes belonging to Spiraeoideae in the classical sense; R = Rosoideae; M = Maloi- deae; P = Prunoideae. The following genera have not found a natural place in one of the eleven (or thirteen after dividing two of them) tribes mentioned above: — the genera composing the tribe Quillajeae in Hutchinson’s classification, a rather hetero- geneous assemblage of Spiraeoid problem cases: Quillaja, Kageneckia, Exochorda, Lindleya, Vauquelinia, Lyonothamnus; — a number of lone genera of uncertain disposition as to tribe and often also subfamily: Holodiscus, Rhodotypos, Kerria, Neviusia, Cercocarpus, Coleogyne, Filipendula, Potaninia, Adenostoma. None of the genera mentioned above occur in Malesia. Morphology — As apparent from the family description, there is variation in many characters of leaves, flowers, and fruits. The presence of a well-developed hypanthium, a probably axial outgrowth from the top of the pedicel surrounding the pistil(s), is about the only character that is common to all Rosaceae. The elaboration of this hypanthium causes much of the variation in flowers and fruits. The plesiomorphic (original) situation is still present in Spiraeoid genera that have a small number (up to 5) multi-ovulate ovaries on the bottom of a cupular hypanthium, the ovaries developing into ventrally dehiscent, dry- walled follicles containing several seeds. Adnation of the ovaries to the inside of the hypanthium, accompanied by a more or less complete fusion of the ovaries with each other, creates the possibility for the evolu- tion of the fleshy, (semi-)inferior fruits that are typical for Maloideae. In this group the exocarp of the inferior fruit is certainly hypanthial, the endocarp (membranous to woody) is certainly carpellary, the more or less fleshy mesocarp may be either or both. In the descriptions in this treatment the terms exocarp, mesocarp, and endocarp are used in 230 Flora Malesiana ser. I, Vol. 11 (2) (1993) their topographical sense, for superior as well as inferior fruits and thus not implying a carpellary origin. Another line of evolution is the change of dry, multi-seeded follicles into dry, 1-seeded achenes, that later may develop a fleshy fruitwall and become drupaceous. Examples are manifold in Rosoideae (e.g. Rubus with many pistils per flower) and all Prunoideae have drupes too, with one pistil per flower. Vegetative Anatomy — Although many papers have been published on the leaf anat- omy of the Rosaceae, little is known of the anatomy of the tropical representatives, especi- ally of the Malesian species. The situation for wood anatomy is much better with recent comprehensive studies by Zhang (1992), Zhang & Baas (1992) and Zhang et al. (1992). The following is a concise summary for the Malesian representatives (wild and cultivated) of data surveyed more extensively in Metcalfe & Chalk (1950) and the above-mentioned wood anatomical studies, amplified with scattered data from the other papers cited below. Leaf anatomy. Trichomes if present usually unicellular, but tufted or stellate hairs occur in Potentilla p.p. and Rubus p.p.; stalked capitate glands have been recorded in Alchemilla, Fragaria, Potentilla, Prunus, Rosa, Rubus, and Sanguisorba. Epidermal cells of lower epidermis sometimes papillate. Extrafloral nectaries present on the petiole and various parts of the leaf blade of Prunus p.p.; leaf teeth glandular or hydathodal in species of Alchemilla, Fragaria, Prunus, Pyrus, Rubus, Sanguisorba, and Spiraea. Stomata almost always confined to the lower leaf sur- face, usually anomocytic, but cyclocytic, staurocytic, tetracytic, and actinocytic types may also occur (Lu et al. 1991). Upper epidermal cells often (partly) mucilaginous. A hypoder- mis is differentiated in some species of Prunus (‘Pygeum’) and Rubus (section Micrantho- batus). Mesophyll dorsiventral. Vascular bundles of minor veins with or without scleren- chyma in bundle sheath, only rarely vertically transcurrent. Vascular system of midrib and petiole ranging from a single collateral bundle to more complex, open or closed systems. Nodes usually trilacunar, but 5-, 7-, and 9-lacunar nodes recorded in Rubus (Kato 1966, 1967). Crystals solitary and/or clustered. Tanniferous cells common. Mucilage idioblasts present in the mesophyll of some species. Wood anatomy. Growth rings faint or absent. Vessels diffuse, typically mostly soli- tary, but radial vessel multiples common in Prunus s.1., and vessel clusters fairly common in Rubus. Vessel frequency and diameter very variable depending on plant habit and ecol- ogy. Intervessel pits, nonvestured, alternate, ranging from minute to large (2-12 um); vessel_ray pits usually similar but half-bordered and slightly smaller (but conspicuously smaller in Prunus p.p.). Helical wall thickenings present in Rosa, all Old World Maloi- deae, and Prunus, but usually weakly developed or restricted to vessel element tails in the Malesian species (as far as studied). Gummy contents common in heartwood vessels. Ground tissue fibres typically fibre-tracheids with distinctly bordered pits common in radial and tangential walls, but in Prunus pits mainly confined to the radial walls, and in some species (belonging to the subgenera Amygdalus and Laurocerasus) also much re- duced in size so that fibres tend to the libriform type. Parenchyma typically scarce, scanty paratracheal and apotracheal diffuse; in the Maloideae parenchyma more abundant and sometimes diffuse-in-aggregates. Irregularly zonate parenchyma bands restricted to Prunus Kalkman — Rosaceae 231 p.p. (‘Pygeum’ and subg. Laurocerasus p.p.). Rays 1—3(—4)-seriate in all Maloideae and in Alchemilla (1-seriate) and Potentilla {1—2(—3)-seriate]; in the remaining genera (Prunus and most Spiraeoideae and Rosoideae) rays of two more or less distinct sizes, with the wide rays 3—8(—16)-seriate. Ray composition varying from homocellular in Micromeles and Photinia p.p.; heterocellular with one to several rows of square to upright marginal cells in Prunus and most Malesian Maloideae (Kribs’ heterogeneous II & III); to largely composed of square to upright cells in all shrubby Spiraeoideae and Rosoideae (so-called ‘juvenilistic rays’). Crystals absent, or present as rhomboidal crystals in ray cells (Rosoideae and Spiraeoideae) or in, usually enlarged, chambered axial parenchyma cells (Maloideae); in species of Prunus druses (subg. Amygdalus, Laurocerasus p.Pp., Padus, Prunus, and ‘Pygeum’) may occur as well as, or instead of rhomboidal crystals (the latter are found in subg. Laurocerasus and Padus), in ray and/or axial parenchyma cells. Traumatic gum ducts sometimes present in species of Prunus (except species be- longing to subg. Prunus). The wood anatomical diversity of the Rosaceae lends itself well for microscopic wood identification and a contribution to phylogenetic classification of the family (Zhang 1992). The Spiraeoideae and Rosoideae are wood anatomically fairly heterogeneous but insepar- able: the Maloideae are a very coherent group showing only very limited wood anatomical variation. Prunus, on the other hand is wood anatomically very diverse and several groups, largely coinciding with the present subgeneric boundaries, can be distinguished, lending some support to their treatment as separate genera: Amygdalus, Laurocerasus, Padus, Prunus s.s., and Pygeum. More detailed leaf and wood anatomical studies of the Male- sian taxa will certainly yield much information of taxonomic significance. References: Devadas, C & C.B. Beck, Amer. J. Bot. 59 (1972) 557-567 (nodal anatomy Physocar- pus, Prunus, Rubus, Potentilla, Geum). — Kato, N., J. Jap. Bot. 41 (1966) 101-107 & ibid. 42 (1967) 161-168 (nodal anatomy). — Lersten, N.R. & J.D. Curtis, Can. J. Bot. 55 (1977) 128-132 (trichomes of Rubus, Physocarpus, Fragaria, Potentilla), ibid. 60 (1982) 850-855 (hydathodes and water pores). — Lu, L.T., Z.L. Wang & G. Li, Cathaya 3 (1991) 93-108 (leaf anatomy of Eriobotrya, Photinia, Rhaphio- lepis, Sorbus, Stranvaesia). — Metcalfe, C.R. & L. Chalk, Anatomy of the Dicotyledons 1 (1950) 539- 550. — Morvillez, M.F., Recherches sur l'appareil conducteur foliaire des Rosacées, des Chrysobalana- cées et de Legumineuses. Thesis, Lille (1919). — Schnell, R., G. Cusset & M. Quenum, Rev. Gén. Bot. 70 (1963) 269-342 (extrafloral nectaries). — Simomura, T. & H. Kurokawa, J. Jap. Bot. 26 (1951) 339-343 (leaf anatomy Prunus). — Singh, V. & D.K. Jahn, Curr. Sci. 44 (1975) 63 (stomatal poly- morphism in Prunus). — Zhang, S.-Y., Blumea 37 (1992) 81-158 (systematic wood anatomy of the Rosaceae). — Zhang, S.-Y. & P. Baas, [AWA Bull. n.s. 13 (1992) 21-91 (wood anatomy of Chinese Rosaceae). — Zhang, S.-Y., P. Baas & M. Zandee, [AWA Bull. n.s. 13 (1992) 307-349 (ecological wood anatomy). P. Baas Palynology — Pollen of representatives of some 80 genera of Rosaceae has been stud- ied in greater or less detail (see Tissot 1990). A comprehensive family treatment is not available up to now. Detailed regional accounts are those by Reitsma (1966), Teppner (1966) and Eide (1981) for NW Europe, and those by Hebda et al. (1988a, b, 1990, 1991) for W Canada, and by Naruhashi & Toyoshima (1979) for Japan. Rosaceae is a stenopalynous family. The pollen grains are isopolar, radially symmetric, subspheroidal monads. The polar axis (P) is 10-60 um (mostly 20—35 um), the equato- 232 Flora Malesiana ser.I, Vol. 11 (2) (1993) rial diameter is 10-50 um. The Spiraeeae, Gillenieae, Holodiscus, Filipendula and Ade- nostoma have small pollen (P usually < 20 um). The largest grains occur in Agrimonia and Mespilus (P up to 50—60 pm). The apertural system is generally tricolporate, though di-, tetra-, syn- and pericolporate grains may be occasionally found in tricolporate samples. Small grains have often indistinct or ill-defined endoapertures (colpate, colporoidate). Pollen of Woronowia, a monotypic genus included in Sieversia by Hutchinson (1964), is 5- (or 6-)colporate (Li 1990). Usu- ally the colpi are relatively long, but Polylepis and Cliffortia (Poterieae) have brevicolpate and porate ectoapertures respectively. All Potentilleae have operculate ectoapertures. Oper- cula occur also in several genera of Poterieae (e.g., Acaena, Agrimonia, Sanguisorba), and in a number of Rosa species. In Sanguisorba spp. the opercula are as wide as the mesocolpia, which often led to describing the grains as being 6-colporate. Many rosace- ous pollen grains show ‘pore flaps’, i.e. more or less protruding sexinous extensions of the mesocolpia arching over an endoaperture, and that occasionally may form an equa- torial bridge (Hebda & Chinappa 1990). Exine stratification is usually distinct with light microscopy. The scarce electron micro- graphs published show a columellate infratectal layer. Ornamentation is mostly meridional- ly striate, but much variation is found in the length, height, width and pattern of the muri and lumina, and the size and number of the perforations within the lumina. Agrimonia and Sor- baria pollen is finely transversely striate. Other ornamentation types include rugulate, psilate (e.g. Cotoneaster), perforate-microreticulate (Neillia) and verrucate-scabrate (e.g., Acaena, Alchemilla, Sanguisorba). Many intermediate forms exist. The tribe Poterieae is most di- verse with respect to ornamentation. The scabrate type of Acaena and Polylepis might be associated with wind-pollination. Pollen of these plants is recorded in many diagrams of lake sediments from Colombia and Venezuela (Smit 1978; Salgado-Labouriau 1979). Obviously, Rosaceae pollen offers few possibilities for subdividing the family. The detailed studies of Canadian Rosaceae pollen allow the recognition of many pollen types (Hebda et al. l.c.), but future work must reveal whether these types have any systematic and phylogenetic significance. Very few fossil pollen has been reported, all from the Oli- gocene onwards (Muller 1981). The pollen of Chrysobalanaceae and Rosaceae is readily distinguisable, but the differ- ences are relatively small (Prance 1989). Pollen of the Neuradaceae is clearly different (see Erdtman 1952; Van Zinderen Bakker & Coetzee 1959). References: Eide, F., Grana 20 (1981) 101-118. — Erdtman, G., Polien morphology and plant taxon- omy (1952). — Hebda, R.J. & C.C. Chinnappa, Rev. Palaeobot. Palynol. (1990) 103-108. — Hebda, R.J., C.C. Chinnappa & B.M. Smith, Grana 27 (1988a) 95-113; Can. J. Bot. 66 (1988b) 595-612; ibid. 68 (1990) 1369-1378; ibid. 69 (1991) 2583-2596. — Hutchinson, J., The genera of flowering plants 1 (1964). — Li, C.-L., Chin. J. Bot. 2 (1990) 150-153. — Muller, J., Bot. Review 47 (1981) 1--142. — Naruhashi, N. & Y. Toyoshima, J. Phytogeogr. Tax. 27 (1979) 46-50. — Prance, G.T., in Flora Malesiana I, 10 (1989) 635-678. — Reitsma, T., Acta Bot. Neerl. 15 (1966) 290-307. — Salgado-Labouriau, M.L., Grana 18 (1979) 53-68. — Smit, A., Rev. Palaeobot. Palynol. 25 (1978) 393-398. — Teppner, H., Phyton (1966) 224-238. — Tissot, C., Sixth bibliographic index to the pollen morphology of Angiosperms (1990). — Zinderen Bakker, E.M. van & J.A. Coetzee, South African pollen grains and spores 3 (1959) 104-200. BW. JM aaendeukinns Kalkman — Rosaceae 233 Phytochemistry — General remarks. Chemistry and chemotaxonomy of Rosaceae were reviewed twice in recent time (Hegnauer 1973, 1990). In these treatises Chrysoba- lanaceae with their characteristic seed oils were included as a subfamily in Rosaceae. A series of papers treating the impact of secondary metabolites, predominantly phenolic compounds, on the classification and infrafamiliar evolution of Rosaceae was published by Challice (1973, 1974, 1981), and a short chemotaxonomic discussion of Rosaceae was included in an essay of Hegnauer (1976). For the present purpose references usually will only be given for papers not cited in one of the forementioned publications. Worldwide, many members of Rosaceae are highly esteemed in traditional medicine. Several recent investigations are concerned with rosaceous medicinal crude drugs, including some eastern Asiatic ones; they will be mentioned in the present phytochemical summary. Leaf phenolics are often heavily overrated in chemotaxonomic discussions (Hegnauer 1990: 373-375). They are doubtlessly valuable characters at generic and lower levels, but they should be used with extreme care only at higher hierarchic levels; tannins and iso- flavonoids are perhaps the taxonomically most promising phenol classes at suprageneric levels. Cyanogenic glycosides. Since more than 160 years amygdalin, the gentiobioside of mandelonitrile, is known from seeds of bitter almonds, and prunasin, a glucoside of mandelonitrile, was prepared in 1895 from amygdalin by cleaving off one molecule of glucose. For a thorough discussion of cyanogenesis and its possible taxonomic meaning in Rosaceae see Fikenscher et al. (1981). Amygdalin has been detected in seeds of many species of Prunus and several genera of Maloideae. In vegetative parts of both taxa amyg- dalin is usually replaced by prunasin. For a long time these two cyanogenic glycosides, which release benzaldehyde on hydrolysis, were considered to be characteristic of the family. This is true, however, only of the genus Prunus with the basic chromosome num- ber (x = 8) and of Maloideae (x = 17). True Rosoideae (x = 7) do not produce cyanogenic compounds. The genera Exochorda, Oemleria and Prinsepia (all with x = 8 and all consid- ered to be Prunoideae) have weakly cyanogenic leaves and twigs with a still unidentified cyanogenic glycoside which is not prunasin. The Spiraeoideae sensu Schulze-Menz 1964 (mostly x = 9) are rather heterogeneous with regard to cyanogenesis. Prunasin is present in leaves of Gillenia trifoliata, Spiraea prunifolia (release of HCN and benzaldehyde; not by other Spiraea-taxa) and Aruncus silvester (also in rootstocks), and Sorbaria-taxa (possibly also C. hamaebatiaria millefolium) produce heterodendrin and aromatic esters of cardiospermin. Kageneckia with x = 17 pro- duces prunasin and resembles Maloideae in this respect. A third type of cyanogenic glycosides, the tyrosine-derived dhurrin, was demonstrated to be present in Cercocarpus and Chamaebatia (both with x = 9), whereas in other ‘rosoid’ taxa with x = 9 hitherto unidentified cyanogenic glycosides, which do not release benzal- dehyde, occur in small amounts: Kerrieae with Coleogyne ramosissima, Kerria japonica, Neviusia alabamensis and Rhodotypos scandens, and Adenostomeae with Adenostoma fasciculatum and sparsifolium. Thus Rosaceae use at least three pathways, the phenylala- nine-route, the tyrosine-route and the leucine-route, for the production of their cyanogenic glycosides. The leucine-pathway is also known from Mimosaceae, Crassulaceae and Sa- pindaceae. 234 Flora Malesiana ser.I, Vol. 11 (2) (1993) Phenolic constituents. All rosaceous plants are accumulators of phenolic and poly- phenolic constituents, but the profiles of phenolics vary widely with plant parts and with taxa. Simple phenolic glycosides like arbutin [Pyrus, Sorbaria, Adenostoma, but not Exo- chorda (Hegnauer 1990: 374)], picein (the glucoside of 4-hydroxyacetophenone; bark of Amelanchier p.p.), gein (vicianoside of eugenol; Geum s.1. p.p.), sweet-tasting dihydro- chalcone glucosides phloridzin, sieboldin and trilobatin [Pyrus and according to Challice (1974) Docynia and Sorbaria, but not Adenostoma] and the glycosidic derivatives of sali- cylic acid monotropitin and spiraein (Filipendula) are taxonomic markers of the genera mentioned or at least of a number of their species. The closely related species Prunus laurocerasus (prunasin) and P. /usitanica (lusitani- coside, the rutinoside of the monophenolic phenylpropanoid chavicol) can easily be dis- cerned by their major leaf constituents mentioned, and the glucosidic phloracetophenone derivative domesticoside was hitherto only isolated from the bark of Prunus domestica. Mahaleboside is a 5-glucosyloxycoumarin of Prunus mahaleb; coumarin, herniarin and a number of 5-hydroxylated and O-methylated coumarins, such as tomentin and fraxinol, seem to be rather characteristic bark constituents of certain Prunus species. Biologically active acylphloroglucinol derivatives occur in flowers of Hagenia abyssinica, which were formerly used as taenifugum (kosotoxin, protokosin), and in the Chinese medicinal plant Agrimonia pilosa (agrimophol and the agrimols A-C). Lignans (Ayres & Loike 1990) were isolated from Maloideae [9’-xyloside and 9’-rham- noside of (+)-lyoniresinol: Sorbus aucuparia, Cotoneaster depressus] and from Prunoideae [prinsepiol, a furofuranoid lignan from Prinsepia utilis and pygeoside, the 9-xyloside of (-)-lyoniresinol from Pygeum acuminatum]. For aucuparin-like biphenyls see under phyto- alexins. Many more genus- or species-characteristic simple phenolic compounds and their gly- cosides could be listed without difficulties, but at higher taxonomic levels (tribes, subfam- ilies, family) hydroxybenzoic acids, hydroxycinnamic acids, flavonoids and tannins are the predominant classes of phenolic constituents. Bate-Smith (1961, 1962, 1965) showed, that p-coumaric and caffeic acid (= 4-hydroxy- and 3,4-dihydroxycinnamic acid), the flavonols kaempferol and quercetin and cyani- din generated from procyanidins (condensed tannins) occur widely in rosaceous leaf hy- drolysates, that trihydroxylation of the B-ring of flavanoid compounds (e.g. myricetin, prodelphinidins) is rare, and that ellagic acid indicating presence of ellagitannins is re- stricted to true Rosoideae (x = 7). Moreover, Bate-Smith already noted incidental pres- ence of flavones (apigenin, luteolin) and 6-hydroxylated flavonoids (quercetagetin) in the family. The screening for rosaceous leaf phenolics was much extended by Challice. He showed general occurrence of 3-caffeoylquinic acid (chlorogenic acid) and restriction of mixtures of dicaffeoylquinic acids, known as isochlorogenic acid, to many genera of Maloideae, inclusive of sections Aria and Aucuparia of Sorbus and of Lindleya. At this point the non- phenolic cinnamic acid should be mentioned. It occurs in large amounts in hydrolysates of several species of Spiraea; originally it is present as cinnamoyl-B-glucopyranose and an acylated derivative, spirarin. Benzoylglucose was isolated from Luetkea pectinata. Challice Kalkman — Rosaceae 235 stressed the taxonomic importance of flavone-C-glycosides within Maloideae, including Dichotomanthes, and the rather sporadic occurrence of these metabolites elsewhere in the family. As already mentioned, the presence of isoflavones in Rosaceae may be taxonomically rewarding, because they possibly indicate affinities with Leguminosae [see for isoflavones the treatment of Mimosaceae in Flora Malesiana 11 (1), p. 19]. The isoflavones genistein, prunetin, biochanin-A and their glucosides prunitrin (prunetin-4’-glucoside), prunetino- side (prunetin-5-glucoside) and biochanin-A-7-glucoside have not yet been traced as leaf constituents; they were isolated from wood and bark of several species of Prunus, from fruit stalks of Prunus avium and P. cerasus and from flowers and fruits of Cotoneaster pannosa and serotina. Many more peculiar flavonoids including flavanones (e. g. Bilia et al. 1991), flavanonols (e.g. Yoshida et al. 1989a), the 8-methoxyflavonols sexangularetin and cor- niculatusin which occur in Dryas octopetala, Cowania mexicana and Purshia glandulosa (all Dryadeae with x = 9), many more O-methylated flavonoids, tricetin, a flavone with a 3’,4’,5-trihydroxylated B-ring (Luetkea pectinata), and even 2-phenoxychromones (Hashidoko et al. 1991a) are produced by Rosaceae. For more information about the multiformity of flavonoid metabolism in the family and its possible taxonomic meaning see Challice (1981) and Hegnauer (1973, 1990). The most conspicuous rosaceous hydroxybenzoic acid is gallic acid (3,4,5-trihydroxybenzoic acid); it will be mentioned under tannins. Tannins (compare also Mimosaceae treatment). Proanthocyanidins (condensed tannins) seem to be more of less ubiquitous in leaves, flowers, fruits, stems and roots of Rosa- ceae; they are accompanied by their monomeric building stones (+)-catechin and (-)-epi- catechin and many (4—8)- and (4—6)-linked catechin dimers and trimers with low tanning activity; the proanthocyanidins with strong tanning action are assumed to be usually tetra- mers and higher oligomers. Double-linked A-type procyanidins are known from the bark of Rhaphiolepis umbel- lata and from Prunus spinosa. (+)-Gallocatechin and prodelphinidins which are rarely present in leaves seem to occur more often in ‘tannin’-fractions of fruits (Maloideae p.p.) and roots (Sanguisorba officinalis, Potentilla erecta). Purshia tridentata and Coleogyne ramosissima yielded about 3% of true condensed tannins (average mol.wt. 13-1400, i.e. tetra- to pentamers) from winter dormant twigs of current season growth; the tannins of the two taxa differed in stereochemistry and biological activity (Clausen et al. 1990). The Purshia tannin was found to have a catechin/epicatechin-ratio of about 55: 45 and to be prefered by snowshoe hares in a choice feeding bioassay to the Coleogyne tannin, which is predominantly based on epicatechin. This shows that the always highly complex con- densed tannin-fractions may have an extremely diverse spectrum of biological activities which depend on hydroxylation patterns and stereochemistry of their building stones, on the nature and stereochemistry of interlinkages in the oligo- and polymers and on degrees of polymerisation. In true Rosoideae (x = 7) the situation is even more complex, because in these plants condensed tannins are accompanied by gallo- and ellagitannins, and because catechins may also be linked with other aromatic metabolites; the pilosanols-A to -C are antimicrobial 236 Flora Malesiana ser.I, Vol. 11 (2) (1993) compounds of Agrimonia pilosa in which C-8 of (-)-epicatechin is combined via a methy- lene group with acylphloroglucide residues (Kasai et al. 1992). In recent times hydroly- sable tannins were thoroughly investigated for a number of medicinally used crude drugs. Examples from Rosaceae are roots of Rosa davurica (Yoshida et al. 1989a, 1991), hips and fresh leaves of Rosa laevigata (Yoshida et al. 1989b), petals of Rosa rugosa (Hatano et al. 1990), petals of ‘apothecary’s rose’ (Eugster & Marki 1991), Alchemillae Folium (mainly Alchemilla xanthochlora), which seems to contain only hydrolysable tannins (Geiger 1991), and Tormentillae Radix which is rich in condensed tannins, but also con- tains ellagitannins (Geiger 1991) and which derives from Potentilla erecta. Root cultures of Sanguisorba officinalis yielded gallic acid, (+)-catechin, (+)-galloca- techin, procyanidin-B3, three gallotannins (2,8%) and the ellagitannins pedunculagin, sanguiin-H6 (up to 5,9%) and sanguiin-H11 (up to 2,3%) and 4,6-hexahydroxydiphe- noylglucose (Ishimaru et al. 1990). Strawberries (Fragaria x ananassa cv. Kent) and raspberries (Rubus idaeus) contain small amounts of ellagitannins, and casuarictin was isolated from strawberries (Daniel et al. 1991). Lamaison et al. (1990) investigated 42 rosaceous taxa representing all four traditional subfamilies for tannin content and observed a range from 1,7 (flowers of Kerria japonica) to 25,1% (roots of Potentilla erecta); they demonstrated that biological activity measured by inhibition of the pancreatic endopeptidase enzyme elastase is not correlated with tannin concentration. Most active tannins were found in flowers and leaves of Alchemilla xantho- chlora, flowers of Filipendula ulmaria, aerial parts of Geum montanum and G. rivale and leaves of Sanguisorba minor, all true Rosoideae with x = 7 and usually containing both condensed tannins and hydrolysable gallo- and ellagitannins. This is another indication for the pluriformity of biological activities of individual tannin components. Finally it should be mentioned that trimethylellagic acid was isolated from rhizomes of Sanguisorba offici- nalis, because ellagic acid methylethers are usually conceived as taxonomic markers of Myrtales; they also occur in Euphorbiaceae. A most recent publication of Okuda et al. (1992) confirms confinement of ellagitannins to Rosoideae with x = 7, including Filipendula (!); moreover, oligomeric ellagitannins are assumed to show a genus-specific distribution: sanguiin-H6 and -H11 in Sanguisorba and Rubus, gemin-A in Geum s.1., agrimoniin in Agrimonia, Fragaria and Potentilla and ru- gosin-D in Filipendula; this publication considers leaf tannins only. For fruits, restriction of ellagitannins to Rosoideae was shown by Foo & Porter (1981). Sugars and hexitols. Saccharose is present in all Rosaceae in appreciable amounts; it is an easily metabolised temporary carbohydrate reserve and is used for the transport of car- bohydrates. In Prunoideae, Maloideae, most Spiraeoideae in traditional circumscription and in Kerrieae and Adenostomeae part of saccharose is replaced by the hexitol sorbitol (= gluci- tol) which got its name from Sorbus aucuparia, one of the first and best sources of this sugar alcohol. Lack of appreciable amounts of sorbitol is a character of the ellagitannin-produc- ing true rosoids inclusive of Dryadeae-Geinae, all with x = 7. In this respect part of Drya- deae sensu Schulze-Menz (1964), i.e. Dryadinae and Cercocarpinae both with x = 9, seem to agree better with true Rosoideae (x =7) than with Spiraeoideae and Kerrieae with x = 9. Waxes and other lipids. Rosaceous cuticular waxes of leaves and fruits (Maloideae) usually are rich in free pentacyclic triterpenic acids. Ursolic and oleanolic acid and a num- Kalkman — Rosaceae 237 ber of hydroxylated derivatives (e.g. maslinic and pomolic acid) were isolated from leaves and pomes of many taxa. Barks of Rosaceae often contain free and/or esterified pentacyclic triterpenic alcohols and ketones; lupeol, betulin, 23-hydroxybetulin (sorbicor- tol-B), taraxerol, alnusenol (= glutinol), friedelanol, alnusenone (= glutinone), friedelin and others were isolated from several taxa. Moreover, barks also contain free and esteri- fied triterpenic acids; examples are pyracrenic acid (betulinic acid-3-coumarate) from Pyra- cantha crenulata and betulinic and 3-epibetulinic acid from Spiraea cantoniensis. Most non- glycosylated bark triterpenes are probably constituents of cork waxes. Besides triterpenes, lipid fractions of all plant parts contain alkanes, alkenes, alkanols, long-chain fatty acids and mixtures of phytosterins. Saponins and pseudosaponins. Saponins are widespread in the family. The sapogenins are usually derivatives of pentacyclic triterpenes, mostly ursolic acid, but sometimes olea- nolic or betulinic acids. Three types of sugar attachment to the sapogenins occur: 3-glyco- sides, esters of the 28-carboxyl group, and the bisdesmosidic saponins which have both linkages. Compounds which have only the ester-linkage were called pseudosaponins by French authors, because their properties are different from those of 3-glycosides (true saponins); tormentol or tormentoside (= rosamultin) is such a pseudosaponin which oc- curs in many true Rosoideae (x =7) and in some Maloideae, but could not be detected neither in Prunoideae and Spiraeoideae nor in Kerrieae; it is the 28-COOH glucose ester of tormentillic (= tormentic) acid (2a,19a-dihydroxyursolic acid) and seems to combine easily with condensed tannins to antibiotically active adducts. As atule saponins are complex mixtures of closely related compounds. Since a long time quillajasaponin is known; it occurs in the bark of Quillaja saponaria which is com- mercially available (Soap bark) and is (was) mainly used as a non-ionic detergent (soap substitute). Presently quillajasaponin is known to be a mixture of bisdesmosidic com- pounds of complex structure which have quillaic acid (= 16-hydroxy-23-oxo-oleanolic acid) as sapogenin and are acylated in the sugar part by two molecules of 3,5-dihydroxy- 6-methyloctanoic acid (a normonoterpenoid compound). The Japanese crude drug Sanguisorbae Radix (= ‘Ziyu’) gathered from Sanguisorba officinalis yielded the 3-arabinoside of pomolic acid (ziyu-glycoside-II) and its bisdes- mosidic derivative ziyu-glycoside-I which has its 28-carboxyl esterified with glucose. In recent time the pseudosaponins of leaves of Japanese Rubus-taxa were investigated thoroughly. The 28-COOH glucose esters of the ursolic acid derivatives acuminatic (= euscaphic) acid, 19-hydroxyasiatic (= 23-hydroxytormentillic) acid and the 3-epimer of 19-hydroxyasiatic acid could be isolated from R. microphyllus (‘Niga-ichigo’; yielded the niga-ichigosides-F1 to -F3), R. trifidus (‘Kaji-ichigo’; yielded the kaji-ichigosides-F1 and -F2), R. koehneanus (niga-ichigoside-F1 and -F2) and R. x medius, a trifidus hy- bride (niga-ichigoside-F2 and kaji-ichigoside-F1). Moreover, roots of the Chinese R. sua- vissimus contain niga-ichigoside-F1 and suavissimoside-R1, which is a 28-COOH glu- cose ester of a derivative of 19-hydroxyasiatic acid (23-CH2OH oxidized to 23-COOH). Niga-ichigoside-Fl was also isolated from Geum japonicum and from hips of Rosa sterilis. Russian Sanguisorba minor s.1., i.e. Poterium lasiocarpum and P. polygamum, yielded a pseudosaponin with caccigenin (20,218,23-trihydroxyoleanolic acid) as sapo- genin. 238 Flora Malesiana ser.I, Vol. 11 (2) (1993) Recent investigations of leaves of Eriobotrya japonica (Liang et al. 1990, De Tommasi et al. 1992), Potentilla fruticosa (Ganenko & Semenov 1989), hips of Rosa davurica (Kuang et al. 1989), fruits of Rubus coreanus, crataegifolius and parvifolius (Ohtani et al. 1990), whole plants of Rubus ellipticus (Pal et al. 1991) and root bark or roots and aerial parts of Sarcopoterium spinosum, Sanguisorba minor s.1. and officinalis, and species of the endemic Canary Island genera Bencomia, Marcetella and Dendriopoterium (Reher et al. 1991; Reher 1991) for pentacyclic triterpenes and their pseudosaponins and saponins yielded taxonomically remarkable results. For instance, Sanguisorba officinalis which only contains the ziyu-glycosides-I and -II and some recently detected other derivatives of pomolic acid (Cheng & Cao 1992) is distinct from all investigated members of subgenus Poterium, Sarcopoterium spinosum and the Canary Island endemics, which all produce 23-hydroxytormentillic acid and its 28-ester glucoside (= niga-ichigoside-F 1). Rubus coreanus is distinct from other eastern Asiatic species of Rubus by having 0.14— 0.25% coreanoside-F1, a dimeric pseudosaponin, in leaves and fruits (Ohtani et al. 1990). Leaves and fruits of Rubus foliolosus do not contain pseudosaponins, but a mixture of goshonosides (see sub diterpenes; Ohtani et al. 1990). The triterpene (and sesquiterpene) glycoside profiles of leaves of Eriobotrya japonica growing in China are different from those of plants growing in Italy (Liang et al. 1990; De Tommasi et al. 1990, 1991, 1992). Rosaceae also produce tetracyclic dammarane- and cucurbitane-type triterpenes. They were detected in leaves and twigs of Cowania mexicana (dammarenediol-II) and Cercocar- pus intricatus (isofouquierol) and in fruits of Purshia tridentata (bitterbrush; the intensely bitter cucurbitacins-D and -I; occur possibly also in other parts of the plant); all three taxa belong to Dryadeae-Cercocarpinae and -Purshiinae with x = 9. Diterpenes. Isoprenoid C29-compounds seem to be rather rare in the family. They oc- cur in Spiraea japonica and koreana as atisane-type diterpene alkaloids (spirasines, spira- mines, and others). In the genus Rubus several eastern Asiatic species were shown to produce large amounts of C29-glycosides in leaves instead of the usual pseudosaponins. The Chinese Rubus suavissimus contains the intensely sweet-tasting kauranoid steviol- 13,18-bisglucoside rubusoside. From R. chingii, a species also occurring in Japan, where it is called ‘Gosho-igicho’, non-sweet labdanoid mono- and bisglucosides, the goshono- sides-F1 to -F5, were isolated; such diterpenes also occur in fruits of the respective spe- cies (Ohtani et al. 1990), whereas their roots contain pseudosaponins (e.g. R. suavis- simus). Seed reserves. Rosaceae store mainly proteins and fatty oils in their seeds; starch is absent. The seed oils belong to a ‘normal’ type with oleic and linoleic acids as main fatty acids; saturated acids (mostly palmitic) usually approximate 10—15%. Some species of temperate regions (Filipendula ulmaria, Sanguisorba minor, Rosa p.p.) have linolenic acid as a third main fatty acid. Reher (1991) found a 18:3/18:2-ratio of 1.4—2.4 in Poten- tilleae and pf 0.4—1.0 in Sanguisorbeae. Some species of Prunus, notably P. africana (= Pygeum africanum), P. mahaleb, spinulosa, undulata and yedoensis and others, devi- ate from the patterns mentioned by having octadeca-9,1 1,13-trienoic (= elaeostearic) acid as a main fatty acid and resemble in this respect Chrysobalanaceae. Miscellaneous. Rosaceae produce and store many more classes of metabolites. Exam- ples are: Kalkman — Rosaceae 239 (a) Non-volatile organic acids, such as the ubiquitous malic, citric and succinic acids, and ascorbic acid (= vitamin C) which is present in large amounts in the hips of many species of Rosa. Isocitric acid which is seldom present in appreciable amounts in plants is stored in leaves of most investigated species of Rubus. (b) Some species with glandular hairs produce essential oils containing mostly mono- and sesquiterpenoid constituents. Such essential oils are also deposited in the wood of certain species of Prunus. The best known ‘volatile oils’ of Rosaceae are rather products of hydrolysis of glycosides than true essential oils and usually consist for over 90% of one or two compounds, e.g. bitter almond oil (benzaldehyde from prunasin and amyg- dalin), methylsalicylate and salicylic aldehyde (from monotropitin and spiraein), eugenol (from gein), chavicol (from lusitanicoside) and Sorbus aucuparia fruit oil, which consists of antibiotically active parasorbic acid, the lactone of 5-hydroxy-2-hexenoic acid (= 2- hexene-5-olide); parasorbic acid is not present as such in the bitter fruits and seeds of Sorbus species of section Aucuparia, but as the glucosidic bitter precursor parasorboside, which is 3-glucopyranosyloxy-5-hexanolide, and seems to be a chemical marker of Sor- bus section Aucuparia. The very expensive true oil of rose is produced from fragrant flowers of several taxa of Rosa and contains predominantly the monoterpenic alcohols citronellol, geraniol and nerol and appreciable amounts of phenylethylalcohol; in fresh young petals these alcohols are present as glycosides. Glycosides of alcoholic mono- and sesquiterpenes seem to be rather common in the family; some recent examples are leaves of Eriobotrya japonica (De Tommasi et al. 1990, 1992) and leaves of Spiraea cantoniensis (Takeda et al. 1990). (c) Characteristic constituents of fruit aromas, such as raspberry (Rubus idaeus), straw- berry (Fragaria), quince (Cydonia oblonga), apples (Malus) and cherries (Prunus); in fresh fruits glycosidic precursors may be present: (d) Nitrogen-containing constituents like the proline derivatives of Malus and other Maloideae and the amines present in the foetid flower smell of some rosaceous taxa (€. g. Crataegus p.p., Sorbus p.p.). According to Strack (1990) Rosaceae are characterized by the production of N,N,N-tricoumaroylspermidine in flowers, especially in their androecia. From a taxonomic point of view metabolites mentioned sub (a) to (d) are unimportant at suprageneric levels, if the triacyl derivatives of spermidine are excluded. It seems there- fore to be more rewarding to finish this short chemical survey with a few remarks on phy- toalexins and recent publications on Prunus constituents. Phytoalexins. Phytoalexins are antibiotically active compounds produced by plants after stimulation by infections or similar stresses. Phytoalexins became known from Ro- saceae only recently. The chemical nature of phytoalexins produced by a taxon depends to some extent on the triggering agents and the plant parts. Nevertheless some taxonomically interesting trends can be discerned in Rosaceae. Maloideae tend to produce aromatic phy- toalexins based on the biphenyl and benzofuran skeleton. The biphenyls aucuparin, 4’- methoxyaucuparin and rhaphiolepin are produced in infected sapwood or bark of Malus pumila, Eriobotrya japonica and in stressed leaves of Eriobotrya japonica and Rhaphio- lepis umbellata, and the biogenetically related benzofurans a-,B- and y-pyrufurans, coto- nefuran and eriobofuran were extracted from infected sapwood of Pyrus communis and 240 Flora Malesiana ser.I, Vol. 11 (2) (1993) Cotoneaster lacteus and from diseased leaves of Eriobotrya japonica (Kemp & Burden 1986; Watanabe et al. 1982, 1990; Miyakado et al. 1985). Phenylpropanoid sapwood phytoalexins are the coumarin scopoletin of Prunus domestica and the lignan iso-olivil from Prunus jamasakura (Kemp & Burden 1986). Benzoic acid was shown to be the anti- fungal compound produced after infection by Nectria galligena in apples of cv. Bramley’s Seedling; it can prevent or retard fruit rotting during storage (Swinburne 1973). Usually the production of phytoalexins is connected with necrosis of attacked cells. Sapwood phytoalexins are comparable to compounds present in heartwoods which only contain dead wood parenchyma cells. Aucuparin, for instance, occurs in heartwood of all investigated species of Sorbus sect. Aucuparia. Moreover, what is known as a phytoalexin from one plant part may be a normal constituent of perfectly healthy tissues of another part of the same plant or of other plants, e.g. the coumarin scopoletin, the lignan iso-olivil and benzoic acid. Lastly, the definition of phytoalexin is rather vague; small amounts of a given phytoalexin of a given taxon may be present in its healthy tissues. Therefore trigger- ing of intensified synthesis of a compound by stress is included by some authors in the phytoalexin concept. Scopoletin in bark of Prunus domestica, coumarin and biogenetically related compounds in leaves of Prunus mahaleb and glycosides of gentisic acid in wood and bark of Prunus yedoensis are examples of ‘phytoalexins’ which are already present in small amounts in healthy plant parts. Chondrostereum purpureum infection induced not only synthesis of aucuparin, but also of 2-dehydrotormentillic acid in sapwood of Malus pumila (Kemp et al. 1985). Isoprenoid phytoalexin-like compounds were also isolated from damaged leaves of Rosa rugosa (Hashidoko et al. 1989); they were shown to be watersoluble sesquiterpenes with the carotane skeleton and named rugosal-A (strongly fungitoxic) and rugosic acid-A (scarcely fungitoxic); both are monohydroxy-endoperoxides which bear an aldehyde resp. carboxyl group. Later the same authors reported, that leaf tissues contain a labile precur- sor, carota-1,4-dienealdehyde which on autoxidation yields rugosal-A and rugosic acid-A (1990), and that leaves additionally produce many more carotane type sesquiterpenes to- gether with acaranoid and bisabolanoid oxigenated Cj5-compounds (1991b, c). Finally Hashidoko et al. (1992) observed that rugosal-A and rugosic acid-A are present in the exudate of the glandular trichomes of Rosa rugosa leaves. The last mentioned observa- tions suggest that rugosal-A is not a true phytoalexin, but a compound generated by aut- oxidation from genuinely present precursors. Recent phytochemical investigations with Prunus-taxa: Subg. Prunus — Prunus spinosa: Phenolics of flowers (Kolodziej et al. 1991), fruits (Ramos & Macheix 1990) and branches (Crespo Ibizar et al. 1992; Gonzalez et al. 1992). Subg. Amygdalus — Prunus davidiana: (+)-Catechin and two flavanone glycosides, prunin and hesperetin-5-glucoside, from stems (Choi et al. 1991). Subg. Cerasus — Prunus avium and cerasus: Comparative investigations of inner bark and seedlings for flavonoids, and detection of prunetin-5-glucoside and tectochrysin-5- glucoside as chemical markers of P. cerasus and of dihydrowogonin-7-glucoside and chrysin-7-glucoside as main flavonoids of P. avium; both species have genistein-5-gluco- side (Geibel et al. 1990, 1991). Prunus serrulata Lindl., a cultigen, yielded 6-caffeoyl- glucopyranoside and 1,6-dicaffeoylglucopyranoside (Ali et al. 1989). Its var. spontanea Kalkman — Rosaceae 241 (= P. jamasakura, the wild Japanese mountain cherry) has bitter fruits with prunasin as main bitter principle (Shimazaki et al. 1991); catechins, sakuranetin-5-glycosides, and the lignanoid compounds sakuraresinol, dihydrobuddlenol-B and racemic lyoniresinol were isolated from its bark (Yoshinari et al. 1990). Fruits of P. maximowiczii are also bitter; they contain bitter tetra- to hexaacylsucroses (acyl = one paracoumaroy] + three to five acetyl residues), epicatechin and a little mandelic acid, but no prunasin (Shimazaki et al. 199 1%. Subg. Padus — Fresh bark of P. buergeri yielded no flavonoids, but mono- and bi- acylated glucopyranoses (caffeic and paracoumaric acid), the 6-caffeoylglucoside of me- valonolactone, and a little grayanin (Shimomura et al. 1988, 1989a). Bark of P. grayana is also free of flavonoids, but contains several caffeic, coumaric and 3,4,5-trimethoxy- benzoic acid esters of glucopyranose, the grayanosides-A and -B and the strongly bitter grayanin which is prunasin with OH-6 of its glucose acylated by caffeic acid (Shimomura et al. 1987). Heartwood of P. grayana yielded taxifolin, dehydrodicatechin-A, the salicin derivatives populin, henryoside and pruyanaside-B and the complex salicylic acid deriva- tives virgaureoside and pruyanaside-A (Shimomura et al. 1989b). Leaf wax of P. grayana contains the antioxidative prunusols (Osawa et al. 1991). The bitter barks of P. padus and P. ssiori contain catechins, bitter prunasin and bitter lignanxylosides lyoniside and ssiori- side, and schisandriside in P. ssiori and prupaside in P. padus; moreover, P. ssiori yielded syringin and glucosyringic acid, and P. padus melilotoside and bitter tetra- and penta- acylated sucroses (Yoshinari et al. 1989, 1990). Subg. Laurocerasus — From green fruits of P. laurocerasus Weinges et al. (1991) iso- lated the primveroside of benzylalcohol. The Prunus-investigations just mentioned demonstrate clearly the enormous intrage- neric variation of phenolic profiles and suggest once again that most phenolics are taxo- nomically useful above all at species, section and genus levels. Summarizing it can be stated that secondary metabolites and seed oils can help a lot to arrive at a satisfying infrafamiliar classification, but procure only vague indications con- cerning family affinities [Chrysobalanaceae, Crassulaceae (phenolics, leucine-derived cya- nogenic compounds), Leguminosae, Sapindaceae]. References and Remarks: Ali, A.A., et al., Pharmazie 44 (1989) 734-735. — Ayres, D.C. & J.D. Loike, Lignans: chemical, biological and clinical properties, Cambridge Univ. Press (1990). — Bate- Smith, E.C., J. Linn. Soc., Bot. 58 (no 370) (1961) 39-54 (Potentilla, Prunus); ibid. 58 (no 371) (1962) 95-173 (Rosaceae: 129-131); Phytochemistry 4 (1965) 535-599 (Quillajeae). — Bilia, A.R., et al., Phytochemistry 30 (1991) 3830-3831 (hexaacetylpyracanthoside). — Challice, J.S., Phytochemistry 12 (1973) 1095-1101 (Pomoideae); Bot. J. Linn. Soc. 69 (1974) 239-259 (Pomoideae); Preslia (Praha) 53 (1981) 283-304 (Maloideae). — Cheng, Dong-Liang & Cao, Xiao-Ping, Phytochemistry 31 (1992) 1317-1320. — Choi, J.S., et al., J. Nat. Prod. 54 (1991) 218-224; Planta Medica 57 (1991) 208-211. — Clausen, T.P., et al., J. Chem. Ecol. 16 (1990) 2381-2392. — Crespo Ibizar, A., et al., J. Nat. Prod. 55 (1992) 450-454. — Daniel, E.M., et al., J. Nat. Prod. 54 (1991) 946-952 (release of ellagic acid from ellagitannins in vegetable food may be a factor preventing carcinogenesis; this is another example showing the diversity of biological activities of vegetable tannins, compare Clausen et al. (1990); Lamai- son et al. (1990)). — De Tommasi, N., et al., J. Nat. Prod. 53 (1990) 810-815 (sesquiterpene glycosides); see also Planta Medica 57 (1991) 414—416 (hypoglycemic effect of leaf sesquiterpenes and triterpenes); J. Nat. Prod. 55 (1992) 1025-1032, 1067-1073 (new sesquiterpenic glycosides and esterified triterpenes). — 242 Flora Malesiana ser.I, Vol. 11 (2) (1993) Eugster, C.H. & E. Marki-Fischer, Angew. Chem. Int. Ed. Engl. 30 (1991) 654—672 (chemistry of rose pigments). — Fikenscher, L.H., et al., Planta Medica 41 (1981) 313-327. — Foo, L.Y. & L.J. Porter, J. Sci. Food Agric. 32 (1981) 711-716 (structure of tannins). — Ganenko, T.V. & A.A. Semenov, Khim. Prirod. Soedin. (1989) 856 (isolation of acids). — Geibel, M., et al., Phytochemistry 29 (1990) 1351— 1353; 30 (1991) 1519-1521. — Geiger, C., Ellagitannine aus Alchemilla xanthochlora Rothmaler und Potentilla erecta (L.) Raeuschel, Thesis Albert Ludwigs-Univ. Freiburg im Breisgau (1991) (see also C. Geiger & H. Rimpler, Planta Medica 56 (1990) 585). — Gonzdlez, A.G., et al., Phytochemistry 31 (1992) 1432-1434 (A-type procyanidins). — Hashidoko, Y., et al., Phytochemistry 28 (1989) 425-430 (antimicrobial sesquiterpene from Rosa rugosa leaves); ibid. 29 (1990) 867-872; ibid. 30 (199 1a) 3837- 3838 (flavone from leaves of Rosa rugosa); Z. fiir Naturforsch. 46c (1991b) 349-356, 357-363 (six bis- aborosaols from Rosa rugosa); Phytochemistry 30 (1991c) 3729-3739; Agric. Biol. Chem. 55 (1991c) 1049-1053 (rugosic acids); Phytochemistry 31 (1992) 779-782 (rugosal and related sesquiterpenes). — Hatano, T., et al., Chem. Pharm. Bull. 38 (1990) 3308-3313, 3341-3346 (trigalloylglucoses and ellagi- tannins). — Hegnauer, R., Chemotaxonomie der Pflanzen 6 (1973) 84-130, 727-730, 784; ibid. 9 (1990) 369-405; Nova Acta Leopoldina, Suppl. no 7 (1976) 45-76; Rosaceae: 57-62 (accumulation of second- ary products and its significance for biological systematics). — Ishimaru, K., et al., Phytochemistry 29 (1990) 3827-3830. — Kasai, S., et al., Phytochemistry 31 (1992) 787-789. — Kemp, M.S. & R.S. Burden, Phytochemistry 25 (1986) 1261-1269 (phytoalexins and stress metabolites in sapwood; refer- ences). — Kemp, M.S., et al., J. Chem. Res. (S) (1985) 154-155. — Kolodziej, H., et al., Phyto- chemistry 30 (1991) 2041-2047 (5 A-type proanthocyanidins). — Kuang, Hai-Xue, et al., Chem. Pharm. Bull. 37 (1989) 2232-2233 (Chinese crude drug ‘Cimeiguo’ yielded 9 triterpenic acids). — Lamaison, J.L., et al., Ann. Pharm. Franc. 48 (1990) 335-340 (tannin contents and inhibitory activity). — Liang, Zhou-Zhong, et al., Planta Medica 56 (1990) 330-332. — Miyakado, M., et al., J. Pesticide Sci. 10 (1985) 101-106 (eriobofuran). — Ohtani, K., et al., Phytochemistry 29 (1990) 3275-3280 (dried fruits = Korean crude drug ‘Bog-bun-ja’). — Okuda, T., et al., Phytochemistry 31 (1992) 3091-3096. — Osawa, T., et al., Agric. Biol. Chem. 55 (1991) 1727-1731 (prunusols). — Pal, R., et al., Indian J. Chem. 30B (1991) 292-293 (saponins). — Ramos, T. & J.J. Macheix, Plantes Méd. Phytothérapie 24 (1990) 14-20 (anthocyanins, caffeoylquinic acids and quercetin glycosides). — Reher, G., Planta Medica 57 (1991), Sonderheft A76—77 (triterpenoid and fatty acid pattern of several genera of Rosoideae). — Reher, G., et al., Planta Medica 57 (1991) 506 (triterpenoids). — Schulze-Menz, G.K., Rosaceae, in A. Engler, Syllabus des Pflanzenfamilien 2, 13th ed. by H. Melchior (1964) 209-218. — Shimazaki, N., et al., Phyto- chemistry 30 (1991) 1475-1480. — Shimomura, H., et al., Phytochemistry 26 (1987) 249-251, 2363- 2366; ibid. 27 (1988) 641-644; Chem. Pharm. Bull. 37 (1989a) 829-830; Phytochemistry 28 (1989b) 1499-1502. — Strack, D., Phytochemistry 29 (1990) 2893-2896. — Swinburne, T.R., in R.J.W. Byrde & C.V. Cutting (eds.), Fungal pathogenecity and the plant’s response (1973) 365-382. — Takeda, Y., et al., Phytochemistry 29 (1990) 1591-1593 (monoterpene glucosides). — Watanabe, K.., et al., Agric. Biol. Chem. 46 (1982) 567-568 (aucuparin); ibid. 54 (1990) 1861-1862 (rhaphiolepin). — Weinges, K., et al., Liebig’s Ann. Chem. (1991) 703-705. — Yoshida, T., et al., Phytochemistry 28 (1989a) 2177-2181; Chem. Pharm. Bull. 37 (1989b) 920-924; Phytochemistry 28 (1989b) 2451-2454; ibid. 30 (1991) 2747- 2752. — Yoshinari, K., et al., Chem. Pharm. Bull. 37 (1989) 3301-3303 (lignan xylosides); ibid. 38 (1990) 415-417 (Prunus padus); Phytochemistry 29 (1990) 1675-1678 (Prunus jamasakura). R. Hegnauer Uses — The usefulness of the Rosaceae is mainly to be found in the presence of many edible, and often delicious fruits. See: Eriobotrya, Fragaria, Malus, Prunus, Rubus, Pyrus. Timber hardly enters the world market, but may well be useful on the local scale. Medicinal uses appear to be scarce in the region judged from label data and Malesian literature. These sources may not give a complete picture considering the extensive use made of Rosaceous species in traditional medicine in East Asia (see also the chapter on phytochemistry, p. 233). Kalkman — Rosaceae 243 KEY TO THE GENERA Miceaves simple entire Or lobed .¥.¢354 2 WS Ee = SS TN SS Oy UI 2 . Leaves compound, pinnate, palmate, or 3-foliolate..................... 15 ESRC eed Berets