Anne Lise Schutte

Fire-survival strategy — a character of taxonomic, ecological and evolutionary importance in fynbos legumes

Variation in the fire-survival strategy of the fynbos, legume tribesPodalyrieae andLiparieae was studied, since it is often the only conspicuous difference between morphologically similar taxa. Two main strategies are apparent: sprouters, taxa which are able to sprout from a woody rootstock after fire and non-sprouters, those which only recruit from seed after fire. In fynbos legumes sprouting and non-sprouting taxa differ in their habitat specificity, population densities, relative regional abundance, and in seed germination tempo. Speciation patterns, as inferred from an analysis of the geographical distribution and habitat specificity of the species, are discussed. Problems relating to the use of sprouting versus non-sprouting as a taxonomic character in fynbos legumes are addressed and possible solutions are given.

The major phenolic compounds in the leaves of Cyclopia species (honeybush tea)

Abstract The phenolic compounds of the leaves of Cyclopia species (tribe Podalyrieae) are of both chemotaxonomic and commercial interest, as the leaves are used to brew a herbal drink known as honeybush tea. Despite the commercial importance of Cyclopia, virtually nothing was known about the chemistry of the leaves prior to the present work. Methanolic extracts from leaves of 22 species were screened for the presence and distribution of phenolic compounds. Three major constituents of the leaves were identified as mangiferin (a xanthone) and glycosides of the flavanones hesperitin and isosakuranetin. The combination of these three compounds is a unique character for Cyclopia, as none of them are present in any of the other genera of the tribes Podalyrieae and Liparieae. Various combinations of the three compounds occur in the different infrageneric groups, but the species are remarkably similar. These results are thus of chemotaxonomic significance at the generic rather than infrageneric level.

Evolutionary relationships in thePodalyrieae andLiparieae (Fabaceae) based on morphological, cytological, and chemical evidence

Taxonomic relationships amongst the genera of the southern African tribesPodalyrieae andLiparieae are discussed. Data gained from morphological, cytological and chemical investigations are analyzed cladistically to determine relationships. The genusCadia (tribeSophoreae) is included in the investigation to establish whether it should be transferred to thePodalyrieae. The results clearly indicate that thePodalyrieae andLiparieae are monophyletic and that they should be united, but thatHypocalyptus andCadia should be excluded. Within the monophyletic group, there are two distinct subclades each supported by three apomorphies. The results also show that there is a strong sister relationship betweenAmphithalea andCoelidium. In the taxonomic treatment theLiparieae are placed into synonymy under thePodalyrieae and two subtribes are recognized. A key to the genera in the tribe is given, followed by a synopsis of the genera.

Chemotaxonomic value of alkaloids in the genus Dichilus

Two bicyclic piperidyl alkaloids (ammodendrine and smipine) and a quinolizidine alkaloid (thermopsine) were identified as major alkaloids of the genus Dichlus. Other piperidyl alkaloids positively identified were bipyridyl, N-methylammodendrineN-acetylhystrine, l-acetyl-l,2,3,4-tetrahydropyridine and piperidinone. The latter two compounds have not been previously reported from the Leguminosae. In the tribe Crotalarieae, the dominance of piperidyl alkaloids appears to be a unique chemotaxonomic character for Dichlus. Our results strongly support the present circumscription and also the isolated position of the genus. An affinity with Melolobium rather than Lebeckia is suggested.

Observations on the occurrence and distribution of alkaloids in some genera and species of the tribe Crotalarieae (Fabaceae)

A general survey of alkaloids in the tribe Crotalarieae showed that useful chemotaxonomic data may be obtained. Although the pyrrolizidine alkaloids of Crotalaria L. have been studied in considerable detail, virtually nothing is known for the other, predominantly southern African genera. The genera Dichilus DC., Lebeckia Thunb., Melolobium Eckl. & Zeyh. and Polhillia Stirton were found to contain several alkaloids in sufficient quantities to warrant further investigation. The major alkaloids (presumably all of the quinolizidine type) seem to be characteristic for each of these genera. Aspalathus L., Buchenroedera Eckl. & Zeyh., Lotononis (DC.) Eckl. & Zeyh., Pearsonia Duemmer, Rafnia Thunb. and Wiborgia Thunb. appear to produce much smaller amounts. The possibility of obtaining useful information from these genera seems limited in view of the large quantities of plant material required to enable proper identification of the compounds. Methods of extraction and detection are described. The R f values of the major alkaloids are given for the different thin-layer chromatographic systems used. Some preliminary identifications were confirmed by mass spectrometry.

Taxonomic Relationships amongst Some Genera of Leguminosae Tribe Crotalarieae and Argyrolobium (Genisteae)

Taxonomic relationships amongst the predominantly southern African genera Dichilus DC., Lebeckia Thunb., Melolobium Eckl. & Zeyh., Polhillia Stirton and Wiborgia Thunb. of the tribe Crotalarieae and Argyrolobium Eckl. & Zeyh. of the tribe Genisteae have been studied. The most obvious similarity between these genera is the fusion of the calyx lobes into a trifid lower lip, or at least a tendency towards such a fusion. The morphology, chromosome numbers and alkaloids of some species and the geographical distribution of the genera have been investigated in an attempt to reach a better understanding of relationships. A tentative phylogenetic tree is presented which indicates that Argyrolobium is the sister group of Polhillia and that it should be referred to the tribe Crotalarieae. The concept of the genus Polhillia is broadened to include two anomalous species of Argyrolobium and Melolobium. Diagnostic characters of the genera are illustrated and discussed.

A Reappraisal of the Generic Status of Liparia and Priestleya (Fabaceae)

The generic delimitation of Liparia L. and Priestleya DC. (Fabaceae, tribe Liparieae) is reevaluated. Traditionally the shape and size of the carinal lobe of the calyx, the shape of the keel petals and the number of flowers per inflorescence were used as diagnostic characters. A study of morphological and alkaloid variation indicates that the two genera run much into one another. Differences in inflorescence and floral structure can be attributed to adaptations to different pollination strategies. The unique combination of major alkaloids occurring in both genera also suggests that they are congeneric. Priestleya is therefore placed into synonymy under Liparia. The nomenclature, synonymy and typification of the genus and the 14 species recognized are presented, and some necessary new combinations are made.

The Tribal Position of Hypocalyptus Thunberg (Fabaceae)

The tribal position of Hypocalyptus Thunberg in the Fabaceae subfamily Papilionoideae is investigated. A phylogenetic analysis in which the Australian Bossiaeeae and Mirbelieae, African Podalyrieae and Crotalarieae, South American Sophoreae and Millettieae, and northern temperate Thermopsideae and Genisteae are included, indicates that there is no direct relationship between Hypocalyptus and any of the tribes. It is therefore proposed that the monotypic subtribe Hypocalyptinae Yakovlev be raised to tribal level as Hypocalypteae (Yakovlev) A. L. Schutte. A description of the tribe is presented, followed by an enumeration of the three species recognized. Hypocalyptus Thunberg is a genus of papilionoid legumes confined to the Cape Floristic Region of South Africa. The genus is easily recognized by its trifoliolate leaves, magenta-pink flowers with a yellow nectar guide, intrusive calyx base, and fused stamens. It comprises three distinct species, which are restricted to the fynbos vegetation of the Western and Eastern Cape Provinces (Dahlgren, 1972). Despite its well-defined generic circumscription, the tribal affinities of Hypocalyptus have been uncertain ever since Benthams (1837, 1839) fundamental classification of the subfamily was published. This is clearly reflected in the number of times the genus has been transferred from one tribe or subtribe to another. Bentham (1837, 1839, 1844) placed the genus in the tribe Loteae subtribe Genistinae, which he later changed to the tribe Genisteae subtribe Cytisinae (Bentham, 1865). Harvey (1862) retained the genus in the Genisteae, recognizing no subtribes. In 1964 Hutchinson raised the Cytisinae to tribal level and allocated Hypocalyptus to the Cytiseae. Polhill (1976, 1981f) transferred the genus to the tribe Liparieae, which Yakovlev (1991) subdivided into the subtribes Lipariinae and Hypocalyptinae. He placed Hypocalyptus in the monotypic Hypocalyptinae. The problem regarding the tribal position of Hypocalyptus is thus quite evident. This has, in fact, also been pointed out by several authors (Dahlgren, 1972; Polhill, 1976, 1981e, 1994; Goldblatt, 1981; Bell et al., 1978; Van Wyk et al., 1994; Van Wyk & Schutte, 1995; Schutte & Van Wyk, 1997). As part of a taxonomic study of the tribes Podalyrieae and Liparieae, Schutte (1995) investigated the position of Hypocalyptus within the tribes. Morphological, anatomical, cytological, and chemical characters were examined and analyzed cladistically to determine interand infratribal relationships. The results clearly indicated that Hypocalyptus is misplaced in the Liparieae and should be excluded, while the Podalyrieae and remainder of the Liparieae are monophyletic and should be united (Schutte, 1995; Schutte & Van Wyk, 1997). Hypocalyptus deviates from the Podalyrieae (including the Liparieae) in no less than nine critical characters (Table 1). Of particular interest are the micromorphological characters, i.e., floral pigmentation (Van Wyk et al., 1994); the accumulation of canavanine in the seed (Bell et al., 1978); a chromosome base number of x = 10 (Goldblatt, 1981; Van Wyk & Schutte, 1995); and ephemeral antipodals in the female gametophyte (Schutte, 1997). These, in addition to the five macromorphological characters of the stamens, seed, and pods, indicate unambiguously that Hypocalyptus does not fit in the Podalyrieae. The aim of this paper is to examine the phylogenetic position of Hypocalyptus within the subfamily Papilionoideae. To this end we scanned the literature and selected all the tribes which are assumed to, or have previously been suggested to, be possibly related to the genus. Based on these criteria, the Australian Bossiaeeae and Mirbelieae, African Podalyrieae and Crotalarieae, northern temperate Genisteae and Thermopsideae, and South American Sophoreae and Millettieae were chosen as outgroups (see, e.g., Dahlgren, 1972; Polhill, 1976, 1981a, 1981g, 1994; Crisp & Weston, 1987). It is important to note that the primary objective of this study is to establish whether Hypocalyptus is directly related to any of the tribes, and not to analyze the relationships among the different tribes. NovoN 8: 178-182. 1998. This content downloaded from 157.55.39.177 on Sat, 19 Nov 2016 04:23:38 UTC All use subject to http://about.jstor.org/terms Volume 8, Number 2 1998 Schutte & van Wyk Tribal Position of Hypocalyptus 179 Table 1. Taxonomic differences between Hypocalyptus and the tribe Podalyrieae. Character Podalyrieae Hypocalyptus 1. Stamen fusion free to open sheath closed tube 2. Seed aril shape interrupted at the micropylar end continuous around the hilum 3. Micropyle type punctate y-shaped 4. Micropyle position within hilar region outside hilar region 5. Pods sessile stipitate 6. Chromosome base number x = 9 x = 10 7. Antipodals persistent ephemeral 8. Floral pigments esters of cyanidin-3-glucoside malvidin-3-glucoside 9. Canavanine absent present ANALYSIS AND RESULTS The data matrix, characters, and character states used for the tribal analysis are given in Table 2. These have largely been taken from Van Wyk and Schutte (1995), with some additions and alterations, e.g., the inclusion of Hypocalyptus as a separate taxonomic unit, the incorporation of the Millettieae and the omission of the Argyrolobium group (now included in the Genisteae; Van Wyk & Schutte, 1995). Data for the Australian Bossiaeeae and Mirbelieae came mainly from Polhill (1976, 1981c, 1981d) and Crisp and Weston (1987, 1995). Information on the Millettieae and Sophoreae has been taken from Geesink (1981) and Polhill (1981b), respectively. Polhill (1976) and Bisby (1981) were consulted for information on the Genisteae, and Turner (1981) for data on the Thermopsideae. Variation in the characters and polarization of character states are discussed in the references given at the end of each character. Where plesiomorphic and apomorphic states co-occur, the taxon was scored for the plesiomorphic state. Autapomorphies for the taxa were excluded from the analyses, since they serve no purpose as grouping char-

Chemataxonomic value of anthocyanins in the tribe liparieae (Fabaceae)

Abstract The major anthocyanins of the tribe Liparieae have been identified. Pink and purple flowers in Amphithalea and Coelidium contain the acetic acid and coumaric acid esters of cyanidin, closely similar to the pattern reported for the tribe Podalyrieae. The purple flowers of Hypocalyptus, however, differ markedly from all the genera of both tribes; they contain malvidin rather than cyanidin and peonidin; the malvidin pigment is present as the 3-glucoside in fresh flowers, and it is not esterified as in all other purple-flowered genera. The petals of Liparia flowers contain the 3-sophorosides of cyanidin and pelargonidin, while the bracts have the 3-glucosides of cyanidin and peonidin. These results support the idea of a close relationship between the Liparieae and Podalyrieae but also agree with other evidence that Hypocalyptus is only superficially similar to the genera of these two tribes.

Taxonomic relationships in the genus Dichilus (Fabaceae — Crotalarieae)

Taxonomic relationships within the genus Dichilus have been investigated. A phenetic analysis of 25 different localities as OTU’s (five of each species) and 67 characters resulted in five distinct groups, corresponding to the five species that were recognized in a recent synopsis. For cladistic analyses, 15 characters were used and the genera Melolobium, Argyrolobium and Lebeckia were used as outgroups. In a further analysis the characters were intuitively polarized. Current cladistic methodology indicated only one fully resolved cladogram of minimal length, irrespective of the choice of outgroup. This cladogram is also supported by several additional characters and biogeographical evidence and is therefore proposed as the best estimate of the phylogeny of the species of Dichilus.