Primary tabs



Herbs (sometimes saprophytic), shrubs, lianas or trees. Leaves simple, entire, usually spirally arranged, sometimes alternate, (semi)decussate or verticillate, sometimes scale-like or absent. Stipules absent but stem sometimes provided with a pair of glands at the nodes. Inflorescence usually raceme-like and un-branched, (supra- or extra-)axillary and/or terminal, sometimes thyrsoid or fasciculate, rarely flowers solitary. Flowers bisexual, more or less zygomorphous, rarely actinomorphous. Sepals 5, free and quincuncial, or the lower (abaxial) 2 connate, sometimes all connate, subequal or the lateral ones larger and then often wing-like (alae) and petaloid. Petals 3 or 5, free or variously united, occasionally also with the calyx, usually adnate to the base of the staminal tube or the filaments, subequal or more often unequal with the lower petal often keellike and frequently pouched, lobed, or crested. Stamens 2-10, usually 8, filaments usually more or less connate except between the upper stamens, often adnate to the petals; Ovary superior, usually 2-locular but occasionally 1-, 3-, 5-, 7- or 8-locular, sessile or sometimes stipitate; Ovules 1 per cell and subapical, or (in Xanthophyllum) 4-more in a 1-locular, bicarpellate ovary with 2 parietal placentas, anatropous, bitegmic and crassinucellate. Fruit various, a berry, capsule, samara or drupe.


Asia-Tropical: New Guinea present, Australasia, Indo-Malaya present, Solomon Islands present, South Africa present, South America present present
About 15 genera and over 1000 species, widespread in temperate and tropical regions of the world, especially well-developed in South America and South Africa. In Malesia 6 genera, of which Polygala and Securidaca (not in Australia) are cosmopolitan, Xanthophyllum and Salomonia Indo-Australian, Epirixanthes Indo-Malayan. The sixth genus is Eriandra which belongs to the tropical American tribe Moutabeae, of which 3 genera are known in South America; Eriandra occurs in New Guinea and the Solomon Islands and represents a marked example of disjunct, tropical trans-Pacific affinities.


Most species seem to be adapted to cross-pollination, but thusfar actual observations have only been reported for some Polygala species (see there). In Epirixanthes the structure of the flowers seems fit for cross-pollination (see there). Self-pollination, however, has been reported for species in diverse genera, also for those species in which cross-pollination has been reported, or is suspected to be possible. It may be concluded tentatively that self-pollination is an effective second-chance possibility for reproduction in the Polygalaceae.


Corresponding to the diversity in fruit and seed types there is a great variation in dispersal types (VERKERKE, 1985). Especially in Poly gala many dispersal types occur: myrmeco-chory, ornithochory, anemochory, diplochory, epizoochory. Ornithochory also occurs in Dicli-danthera, Carpolobia, Atroxima, and probably in some Xanthophyllum species. Moutabea fruits have endozoochorous dispersal by monkeys (VAN ROOSMALEN, 1985); this may also be true for some Xanthophyllum species. Myrmecochory (or perhaps also anemochory) probably occurs in Bredemeyera, Comesperma, and Epirixanthes; anemochory also in Monnina and Securidaca (VAN ROOSMALEN, l.c.), but in the latter also hydrochory is possible. Epizoochory is the possible means of dispersal for Salomonia.


Recently the morphology and ontogeny of ovules, fruits and seeds have been described (and reviewed) by VERKERKE; of Polygala: VERKERKE & BAUMAN (1980); of Xanthophyllum: VERKERKE (1984); of the remaining genera: VERKERKE (1985). LEINFELLNER (1972) demonstrated that there is no principal difference in the ontogeny of the unilocular, multiovulate ovary of Xanthophyllum and the bi- to octoloculate, uniovulate ovary of the other Polygalaceae. The ontogeny of bi- and trisporangiate anthers of Polygala has been described by CHODAT (1891) and VENKATESH (1956). Stipular outgrowths or nodal glands are pseudostipules in the sense of WEBERLING (VAN DER MEIJDEN, 1982: 3). Contrary to the idealistic opinion of CHODAT (l.c.) (also adopted by HUTCHINSON, 1967) and CRONQUIST (1981), the primitive number of stamens is 8; the presence of 10 stamens (Diclidanthera; and a rare abnormality in Xanthophyllum) is a derived character.


The Polygalaceae exhibit an interesting diversity in their leaf and wood anatomy, which has only fragmentarily been explored, especially for the Malesian genera. Hairs if present are unicellular, or more rarely uniseriate. The lower epidermis is papillate in a number of species. Stomata may be of the anomocytic, paracytic or anisocytic type (all three types occur within the genus Xanthophyllum). An adaxial hypodermis is frequently present in the woody species with coriaceous leaves. The vascular pattern in petiole and midrib ranges from a single collateral bundle to a closed cylinder with accessory bundles (again the whole range of the family is represented in Xanthophyllum). The nodes are unilacunar. Unusual tracheoidal idio-blasts in the leaf mesophyll are characteristic for Xanthophyllum.

The secondary xylem of the trees and climbers is characterised by largely solitary vessels with simple perforations, fibres with distinctly bordered pits and heterocellular rays which are usually narrow (1-2-seriate), but may be much wider in the Moutabeae (e.g. Securidaca). Axial parenchyma is mainly paratracheal in Polygala, apotracheally diffuse and diffuse-in-aggregates plus vasicentric to loosely aliform in Securidaca, and apotracheally banded plus vasicentric in Xanthophyllum. Included phloem occurs in the wood of Securidaca (and other Moutabeae).

Despite the anatomical distinctness of Xanthophyllum (mainly through its tracheoidal idioblasts), vegetative anatomy clearly witnesses affinity with other Polygalaceae (epidermal characters, overall leaf histology, solitary vessels and fibre type in the wood; the parenchyma distribution in some Xanthophyllum species is reminiscent of that of Securidaca). In its wood anatomy Xanthophyllum also recalls Trigoniaceae, especially Trigoniastrum.


CRONQUIST'S circumscription of the order Polygalales (1981: 763) reflects the general opinion of systematists; next to the Polygalaceae (with Xanthophyllum as a separate family) it includes the Trigoniaceae, Vochysiaceae, Malpighiaceae as well as the Tremandraceae and the Krameriaceae. Wood anatomical evidence (BRIDGWATER & BAAS, 1982) supports a close affinity of Polygalaceae, Xanthophyllum and Trigoniaceae, but not with the other families. VAN DER MEIJDEN (1982) found no arguments to include Tremandraceae and Krameriaceae. Although there seemed to be little evidence for the inclusion of Krameriaceae in the order (SIMPSON & SKVARLA, 1981; SIMPSON, 1982), the first author recently supported the classification of Krameriaceae next to Polygalaceae, mainly based on serological evidence (BUSE-JUNG, 1979).

Splitting off Xanthophyllum as a separate family has been based on incomplete or erroneous knowledge of the genus. Of the three characters mentioned by CRONQUIST (l.c.), two do not hold: filaments are often partly and sometimes halfway connate in Xanthophyllum, and seeds with copious endosperm occur in four of the seven subgenera. Thus Xanthophyllum differs in a single though compound character from other Polygalaceae, viz. in the structure of the ovary (which ontogenetically is largely similar to other Polygalaceae, cf. LEINFELLNER, l.c.); this differs in the reduced septs, in the doubling of the number of ovules per carpel, and in the height of insertion of the ovules. On the other hand Xanthophyllum has a number of striking similarities with other Polygalaceae in the structure of the flowers, as well as in some vegetative characters (the presence of laminar and nodal glands). Also studies on the morphology of ovules, fruits and seeds (VERKERKE, 1984, 1985), of foliar anatomy (DICKISON, 1973) and wood anatomy (BRIDGWATER & BAAS, 1982) do not present arguments to split Xanthophyllum from the Polygalaceae.

Mainly based on differences in floral structure, CHODAT (1891) distinguished three tribes in the family, Polygaleae, Moutabeae and Xanthophylleae. Both the studies of STYER (1977) and VERKERKE (1984, 1985) revealed that the differences between Moutabeae and Polygaleae are unclear; the results of vegetative and of seed anatomy do not correspond with differences in the flowers. Therefore a formal subdivision of the family is not presented.


Chemical characters of the family were summarized and discussed by HEGNAUER (). A supplement will be included in volume 8. Glycosides releasing methyl salicylate on hydrolysis and saponins with triterpenic acids as their sapogenins are rather common in the family, especially in roots. The most characteristic genuine sapogenin of the family seems to be presenegenin, C30H46O7, an acid-labile derivative of oleanolic acid; it was shown to be mainly sapogenin of roots of several members of Bredemeyera, Carpolobia, Polygala and Securidaca. Closely related sapogenins are polygalacic acid from Polygala paenea and bredemolic acid from Bredemeyera floribunda. Mono-, di- and trimethoxycinnamic acids as well as ferulic and sinapic acid occur widely; usually they are esterified with free sugars such as saccharose or with the sugar-part of the bidesmosidic saponins. Leaf flavonoids seem mostly to be glycosides of quercetin and kaemp-ferol. Polygalitol (= aceritol= 1,5-anhydroglucitol) is a characteristic hexitol derivative of roots of many polygalaceous plants; it occurs free and combined with sugars. Seeds usually store mainly proteins and fatty oils, but no starch. There is still little known about the chemistry of polygalaceous seed oils, but very unusual oils were shown to be produced by Monnina emarginata and Polygala virgata. Three groups of secondary metabolites have still to be mentioned notwithstanding the fact that they are known only from a few species of Polygala at present. These are polyhydroxylated xanthones, and naphthalin- and bibenzyl-y-butyrolactone-type lignans such as podophyllotoxin (e.g. Polygala polygama) and suchilactone (e.g. Polygala chinensis). Moreover, the Central American Polygala paniculata yielded a number of rutaceous coumarins and a diester of khellactone.

The taxonomic relevance of chemical characters was discussed by HEGNAUER in 1969. It may be added now that the occurrence of isoprenylated coumarins and an obvious total lack of iridoids agree well with the assumption of sapindalean (sensu lato) affinities. Tannins too seem to be totally lacking in Polygalaceae; this, however, does not contradict the just mentioned assumption, because within Sapindales s.l. a strong tendency to replace tannins by other types of secondary metabolites is apparent. Finally it should not be forgotten that several chemical characters may prove valuable in future for infrafamiliar classification. — R. HEGNAUER.