Bot. Bull. Acad. Sin. (2002) 43: 57-62

Park and Backlund — Origin of the cyathium-bearing Euphorbieae

Origin of the cyathium-bearing Euphorbieae (Euphorbiaceae): phylogenetic study based on morphological characters

Ki-Ryong Park^1,* and Anders Backlund²

¹Department of Biology, Kyung-Nam University, Masan 631-701, Korea

²Division of Pharmacognosy, Department of Pharmacy, Uppsala University, BMC-Biomedical center, S-751 23 Uppsala, Sweden

(Received October 6, 2000; Accepted August 24, 2001)

Abstract. A cladistic analysis of the subfamily Euphorbioideae was undertaken to elucidate the origin of the cyathium-bearing Euphorbieae and to provide hypotheses about evolutionary relationships within the subfamily. Twenty-one species representing most of the genera within the study group and three outgroup taxa from the subfamilies Acalyphoideae and Crotonoideae were selected for parsimony analysis. An unweighted parsimony analysis of 24 morphological characters resulted in five equally parsimonious trees with consistency indices of 0.67 and tree lengths of 39 steps. The strict consensus tree supported monophyly of the cyathium-bearing Euphorbieae. The sister group relationships of cyathium bearing Euphorbieae with Maprounea (subtribe Hippomaninae) were supported weakly, and the origin of cyathium is possibly in Hippomaneae, or in the common ancestor of Euphorbieae and remaining taxa of Euphorbioideae plus Acalyphoideae. Within the tribe Euphorbieae, both subtribes Euphorbiinae and Neoguilauminiinae are monophyletic, but the African endemic subtribe Anthosteminae is unresolved. The resulting trees support the monophyly of the tribe Stomatocalyceae while the tribe Hippomaneae does not consistently form a clade.

Keywords: Cyathium; Euphorbieae; Phylogeny.

Introduction

In a recent classification of subfamily Euphorbioideae Boiss., Webster (1975, 1994b) recognized six tribes: Stomatocalyceae (Muell. Arg.) Webster, Hippomaneae A. Juss. ex Spach, Pachystromateae (Pax & Hoffm.) Pax, Hureae Dumortier and Euphorbieae Blume. Among these the cyathium-bearing taxa are restricted to the tribe Euphorbieae. Compared to Webster's system, Pax (1924) recognized only two closely related tribes: Hippomaneae and Euphorbieae, and one remotely related tribe, Gelonieae (Muell. Arg.) Pax, on his phyletic tree. However, Gelonieae, corresponding in part to Webster's tribe Stomatocalyceae, was indicated as closely related to tribe Manihoteae (Muell. Arg.) Pax of the subfamily Crotonoideae Pax in Pax's system.

The highly specialized inflorescence—cyathium—of Euphorbieae, is characterized by a central female flower and four or five lateral groups of male flowers within the involucre. Although the unique cyathium in Euphorbieae has been considered evidence of a natural group, the origin and nature of this character have been interpreted diversely (Jussieu, 1824; Mueller, 1866; Haber, 1925; Croizat, 1937, 1942; Gilbert, 1994).

Croizat (1937, 1938) pointed out a strong resemblance between Euphorbieae and Hippomaneae except with regard

to the position of a female flower. Accordingly, the Euphorbia-like cyathium results from the alteration of floral axis and the condensation of the axis of male flower in Hippomaneae. This interpretation is also supported by Pax (1924), whose idea of relationships in Euphorbiaceae indicates that Hippomaneae is most closely related to Euphorbieae.

A recent hypothesis, however, does not support the traditional relationships (Pax, 1924; Croizat, 1937, 1938) but rather suggests origin of the cyathium-bearing Euphorbieae from Crotonoideae (Gilbert, 1994). Within the Crotonoideae, Jatropha-like species with a primary female flower and lateral male flowers are considered ancestors of Euphorbieae (Gilbert, 1994). Croizat (1942) and Hurusawa (1954) on the other hand insisted that the cyathium of Euphorbieae should be derived from the pseudanthium of Dalechampia Plumier ex L. in Acalyphoideae Ascherson via Pedilanthus Necker ex Poiteau, but several authors disputed the Dalechampia origin of Euphorbieae (Venkata Rao, 1971; Webster and Webster, 1972; Webster, 1994a).

Pollen morphology indicates that Euphorbia is similar to Stillingia Garden ex L. and Sapium P. Browne of Hippomaneae in having an even margo and a similar sculpturing pattern, but its pollen clearly differs from that of Crotonoideae and Acalypoideae (e.g. Dalechampia) (Webster and Webster, 1972; Webster and Rupert, 1973; Park, 1997). The pollen grains of Crotonoideae are mostly inaperturate with triangular supratectal elements completely absent in Euphorbieae (Park and Lee, 1988; Nowicke, 1994; Webster, 1994b).

*Corresponding author. Tel: 82-55-249-2240; E-mail: park@kyungnam.ac.kr

Botanical Bulletin of Academia Sinica, Vol. 43, 2002

Tribe Euphorbieae, the largest tribe within the Euphorbioideae, is divided into three subtribes: Anthosteminae (Baillon) Webster, Neoguillauminiinae Croizat and Euphorbiinae Hurusawa (Webster, 1975, 1994b). Most traditional hypotheses have placed the Anthosteminae as a basal lineage within the tribe (Pax, 1924; Croizat, 1937). According to the phyletic tree of Croizat (1937), the Australian endemic Calycopeplus Planchon in Neoguillauminiiae is to be regarded as the group most closely related to Euphorbiinae. The Euphorbiinae is the largest and systematically most problematic subtribe of the Euphorbioideae, members of which are recognized by a true cyathium, which lacks the calyx of male and female flowers. Most of the taxonomic problems surrounding Euphorbiinae are due to its large size, a profusion of intergrading and overlapping characters (Sherff, 1940), and classifications made on the base of local flora without comprehensive phylogenetic studies within the complete ranges of the group (Gilbert, 1994; Webster, 1994b; Park, 1996).

The objectives of the present study were: 1) to investigate the origin of the cyathium-bearing taxa within subfamily Euphorbioideae; 2) to investigate the phylogenetic relationships within tribe Euphorbieae; and 3) to evaluate Webster's (1994b) recent classification and tribal circumscriptions within the Euphorbioideae.

Materials and Methods

Twenty-one genera in the subfamily Euphorbioideae as recently circumscribed by Webster (1994b) were included in the analysis (Table 1). Ingroup monophyly of Euphorbioideae, as compared to sister groups Crotonoideae and Acalypoideae, is supported by a number of apomorphic characters such as whitish latex, simple trichomes, apetalous flowers, lack of a disk, absence of vascular bundles in the inner integument, and perforate-reticulate pollen exine (Webster, 1994b; Tokuoka and Tobe, 1995). Based on the relationships of the Euphorbiaceae (Levin and Simpson, 1994), three species from the Crotonoideae and Acalyphoideae were selected as outgroups for rooting the resulting trees. Taxonomic hypotheses based on morphology (Levin and Simpson, 1994), serology (Jensen et al., 1994), anatomy (Rudall, 1987, 1994), and molecular data (Wurdack and Chase, 1999) strongly supported close relationships among the three subfamilies within Euphorbiaceae.

Morphological characters were obtained by examining herbarium specimens and living materials. Twenty-four characters from 24 terminal taxa were coded (Tables 2, 3). Phylogenetic analyses were conducted on an Apple Macintosh 8600/250 with 80 MBRAM available, using PAUP Version 4.0d65 (Swofford, 1999) under the assump

Park and Backlund — Origin of the cyathium-bearing Euphorbieae

The topologies obtained from the analyses before and after the successive reweighting procedure differ on several points, the main one of which is the identification of two major lineages within the Euphorbioideae, one of which includes the outgroup of Acalyphoideae in the latter. These lineages largely correspond to the well supported tribe Euphorbieae on one hand, and to an assembly of other tribes—albeit weakly supported—on the other hand.

In the lineage consisting of several tribes, moderate support is found for a monophyletic Stomatocalyceae, and weak but unambiguous support for the tribe Hippomaneae. In the unweighted analysis, however, the tribe Stomatocalyceae was indicated as the most basal tribe of the subfamily Euphorbioideae, and the tribe Hippomaneae as being paraphyletic with respect to Euphorbieae.

Within the tribe Euphorbieae, both subtribes Euphorbiinae and Neoguilauminiinae are retrieved as monophyletic, and strongly supported by relatively high supporting values within the studied group, but the

tions of Fitch parsimony (Fitch, 1971). The matrix was analysed using 1,000 random taxon addition sequences followed by the efficient "tree bisection reconnection" (TBR) branch swapping algorithm. The characters were then reweighted using the successive approximations weighting method proposed by Farris (1969). Weights were applied according to the rescaled consistency index, and the procedure was repeated until a stable result was obtained. In order to further evaluate the stability of different branches in the obtained trees, a Bremer support analysis (Bremer, 1988; Källersjö et al., 1992; Bremer, 1994) was performed, making use of the computer program "Autodecay PPC 4.01" (Eriksson, 1998) in combination with PAUP in the generalized manner described by Bremer (1994). Furthermore, bootstrap- (Felsenstein, 1985) and jackknife-analyses (Farris et al., 1996) excluding 37% and 50%, respectively, of the characters were also performed using PAUP with settings analogous to those described above.

Results

The first step of the analysis, with unweighted characters, yielded five equally parsimonious trees with a length of 39 steps, a consistency index (CI; Kluge and Farris, 1969) of 0.67 and retention index (RI; Farris, 1989) of 0.90. The strict consensus tree is shown in Figure 1. After successive approximations reweighting a stable result of 18 equally parsimonious trees with a length of 22.84 steps, CI = 0.75 and RI = 0.94 was obtained, the strict consensus of which is shown in Figure 2.

Botanical Bulletin of Academia Sinica, Vol. 43, 2002

Figure 1. Strict consensus tree of six equally most-parsimonious cladograms obtained from phylogenetic analysis of unweighted morphological data (length = 39; CI = 0.67; RI = 0.90). Numbers appearing on the branches indicate, in order, [bootstrap / jacknife with 37% deletion / jackknife with 50% deletion / Bremer support].

Figure 2. Strict consensus tree of 18 equally most parsimonious cladograms obtained from successive approximations reweighted analysis of morphological data (length = 22.84; CI = 0.75; RI = 0.94). Numbers appearing on the branches indicate, in order, [bootstrap / jacknife with 37% deletion / jackknife with 50% deletion / Bremer support].

paleotropical subtribe Anthosteminae was unresolved (Figures 1 and 2). In the unweighted analysis the genus Cubanthus (Boiss.) Millsp. is placed as a sister group of the remaining taxa within the subtribe Euphorbiinae, albeit weakly supported. In the weighted analysis Cubanthus is unresolved with regards to Euphorbiinae.

Discussion

Phylogeny and Classification

The resulting topology from the analysis largely corresponds to the tribal classification of Euphorbioideae proposed by Webster (1994b).

The monophyly of Euphorbieae, the cyathium-bearing group, is particularly well supported by a number of characters. Branch support values for the clade are substantial in all of the measures. Within the Euphorbieae, three subtribes are recently recognized (Webster, 1975, 1994b): subtribe Anthosteminae, including Anthostema and Dichostemma, subtribe Neoguillauminiinae, containing Neoguillauminia and Calycopeplus, and subtribe Euphorbiinae including the remaining genera. However, the delimitation of these three subtribes is controversial. Croizat (1937) separated Dichostemma from Anthosteminae, and placed it within the subtribe Neoguillauminiinae. In the present analysis our results did not show the monophyly of Anthosteminae, as proposed by Webster (1975, 1994b). Neither do they support the separation of Dichostemma and Anthostema (Croizat, 1937). However, our analyses recovered the Neoguillauminiinae and Euphorbiinae clades, as proposed by Webster (1975, 1994b).

The consensus trees of both analyses well support the monophyly of the Stomatocalyceae. However, the rela

tionships between Stomatocalyceae and other genera of the Euphorbioideae were equivocal because the tribe was grouped together with Acalyphoideae in the weighted analysis, or as the sister group of all other Euphorbioideae. A previous phylogenetic analysis of Malesian Hippomaneae, Pimelodendron Hassk. of Stomatocalyceae always segregated from the tribe Hippomaneae and suggested its divergence within the Euphorbioideae (Esser et al., 1997). A recent molecular phylogenetic study (Wurdack and Chase, 1999) suggests that Webster's subfamily Euphorbioideae is not monophyletic and that Stomatocalyceae is excluded from the subfamily. Although previous pollen studies (Punt, 1987; Nowicke, 1994) indicated the naturalness of subfamily Euphorbioideae, pollen characters such as without marginate colpi (character 13), and rugulate sculpturing pattern (character 12) in the tribe Stomatocalyceae suggest a closer relationship with the subfamily Acalyphoideae (Gillespie, 1994), or at least an exclusion from Euphorbioideae (Pax and Hoffmann, 1931). Its habit, colored latex, and oily endosperm (Webster, 1994b) require additional comparison to taxa outside of Euphorbioideae such as Acalyphoideae to determine whether Stomatocalycinae may be part of the tribe Omphaleae or Plukenetieae in Acalyphoideae. A more global parsimony analysis including most of Acalypoideae and Crotonoideae is required to test the monophyly of Webster's subfamily Euphorbioideae.

In the weighted analysis, the New World tribe Hureae is the sister group of Hippomaneae and supports the classification system suggested by Webster (1994b). The two tribes share a presence of glands at the blade base and two other parallel characters. This supports the opinion of Pax (1924) who combined them into a single tribe Hippomaninae, and placed them as sister groups to each other.

Park and Backlund — Origin of the cyathium-bearing Euphorbieae

The Origin of the Cyathium-Bearing Group

In the unweighted analysis, the cyathium-bearing Euphorbieae forms a monophyletic group. It is, albeit weakly supported, nested within Hippomaneae, which is characterized by bisexual inflorescences with basal pistilate and central staminate flowers and inclinate flowers (Esser et al., 1997). The analyses both provide some support for the view that Euphorbieae seems to be evolved from a common ancestor to Maprounea of Hippomaneae rather than from a Jatropha-like ancestor within the subfamily Crotonoideae (Gilbert, 1994). Furthermore, this interpretation does not contradict the previous hypothesis of Webster and Rupert (1973) on the basis of pollen morphology; they suggested that the Euphorbieae are derived from an ancestral form within the Hippomaneae.

In Hippomaneae, the male inflorescence shows two different types of modification: globose spikes and elongate spikes. The elongate spikes of male inflorescence occur in most of the species in Hippomaneae, while the condensed male inflorescence, characteristic of Euphorbieae, is only found in Maprounea of Hippomaneae. Also, insertion of the glands between the clusters of male flowers is well known in two basal clades, Anthosteminae and Neoguillauminiinae of the Euphorbieae, and it only occurs in Maprounea within Hippomaneae. Croizat (1938) suggested that the structure of the inflorescence in the Euphorbieae is not fundamentally different from that of the Hippomaneae, except for the position of the axes, because the cyathium of Euphorbieae has resulted from the condensation of the male axis of Hippomaneae. Gilbert (1994) recently rejected this suggestion and proposed the Crotonoideae origin of Euphorbieae from a Jatropha-like ancestor. The results of our analysis of morphological data are mostly consistent with the hypotheses of Croizat (1938) and Pax (1924).

However, our weighted analysis supports the alternative hypothesis that the cyathium is evolved from a common ancestor of the cyathium-bearing Euphorbieae and from the remaining taxa of Euphorbioideae plus Acalypha australis of the subfamily Acalyphoideae.

In conclusion, Webster's (1994) tribes Euphorbieae and Stomatocalyceae are well supported, but the tribe Hippomaneae does not form a clade consistently. The cyathium-bearing Euphorbieae is a well supported tribe, but whether the tribe originated within Hippomaneae or had an isolated origin from Euphorbioideae cannot be answered clearly. Therefore, a broad phylogenetic analysis including Acalyphoideae and Crotonoideae is needed to resolve the sister group of Euphorbieae.

Acknowledgments. This paper was supported by the Kyungnam University Research Fund in 1996 to K.-R. P. and a Swedish Natural Science Research Council grant to A.B. Helge Axison Johnssons Stiftelse is acknowledged for funding computer equipment to A.B. We furthermore thank the curators of E, K, CANB, BRI, OKL and SKK for loan specimens and use of their facilities, and Drs. G. Webster and S.-P. Hong for reviewing the manuscript.

Literature Cited

Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795-803.

Bremer, K. 1994. Branch support and tree stability. Cladistics 10: 295-304.

Croizat, L. 1937. Notes on Euphorbiaceae, with a new genus and a new subtribe of the Euphorbieae. Philipp. J. Sci. 64: 397-412.

Croizat, L. 1938. Glands of Euphorbiaceae and of Euphorbia. Chron. Bot. IV. 6: 512-514.

Croizat, L. 1942. Peculiarities of the inflorescence in the Euphorbiaceae. Bot. Gaz. 103: 771-779.

Eriksson, T. 1998. AutoDecay ver. 4.0 (Program Distributed by the Author). Bergius Foundation, Royal Swedish Academy of Sciences, Stockholm.

Esser, H.-J., P. van Welzen, and T. Djarwaningsih. 1997. A phylogenetic classification of the Malesian Hippomaneae (Euphorbiaceae). Syst. Bot. 22: 617-628.

Farris, J.S. 1969. A successive approximations approach to character weighting. Syst. Zool. 18: 374-385.

Farris, J.S. 1989. The retention index and the rescaled consistency index. Cladistics 5: 417-419.

Farris, J.S., V.A. Albert, M. Källersjö, D. Lipscomb, and A.G. Kluge. 1996. Parsimony jackknifing outperforms neighbor-joining. Cladistics 12: 99-124.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.

Fitch, W.M. 1971. Toward defining the course of evolution: Minimum change for a specific tree topology. Syst. Zool. 20: 406-416.

Gilbert, M.G. 1994. The relationships of the Euphorbieae (Euphorbiaceae). Ann. Missouri Bot. Gard. 81: 283-288.

Gillespie, L.J. 1994. Pollen morphology and phylogeny of the tribe Plukenetieae (Euphorbiaceae). Ann. Missouri Bot. Gard. 81: 317-348.

Haber, J.M. 1925. The anatomy and the morphology of the flower of Euphorbia. Ann. Bot. 39: 657-707.

Hurusawa, I. 1954. Eine nochmalige Durchsicht des herkommlichen systems der Euphorbiaceen im weiteren Sinne. J. Fac. Sci. Univ. Tokyo, III, Bot. 6: 209-342.

Jensen, U., I. Vogel-Bauer, and M. Nitschke. 1994. Leguminlike proteins and the systematics of the Euphorbiaceae. Ann. Missouri Bot. Gard. 81: 160-179.

Jussieu, A. 1824. De Euphorbiacearum Generibus Medicisque Earumdem Tentamen. Didot, Paris.

Kluge, A.G. and J. S. Farris. 1969. Quantitative phyletics and the evolution of the anurans. Syst. Zool. 18: 1-32.

Källersjö, M., J.S. Farris, A.G. Kluge, and C. Bult. 1992. Skewness and permutaion. Cladistics 8: 275-287.

Levin, G.A. and M.G. Simpson. 1994. Phylogenetic implications of pollen ultrastructure in the Oldfieldioideae (Euphorbiaceae). Ann. Missouri Bot. Gard. 81: 203-238.

Mueller, J. 1866. Euphorbiaceae. In A.P. de Candolle (ed.), Prodromus 15 (2). Masson and Son, Paris, pp. 189-1269.

Nowicke, J.W. 1994. A palynological study of Crotonoideae (Euphorbiaceae). Ann. Missouri Bot. Gard. 81: 245-269.

Botanical Bulletin of Academia Sinica, Vol. 43, 2002

Park, K. 1996. Phylogeny of New World subtribe Euphorbiinae (Euphorbiaceae). Kor. J. Plant Tax. 26: 235-256.

Park, K. 1997. Pollen morphology of Euphorbia subgenus Agaloma section Tithymalopsis and related species (Euphorbiaceae). Grana 36: 11-16.

Park, K. and S. Lee. 1988. A palynotaxonomic study of the Korean Euphorbiaceae. Kor. J. Plant Tax. 18: 69-94.

Pax, F. 1924. Die Phylogenie der Euphorbiaceae. Bot. Jahrb. 59: 129-182.

Pax, F. and K. Hoffmann. 1931. Euphorbiaceae. In A. Engler and K. Plantl (eds.), Die Natürlichen Pflanzenfamilien ed. 2, 19C, Leipzig, pp. 11-233.

Punt, W. 1987. A survey of pollen morphology in Euphorbiaceae with special reference to Phyllanthus. Bot. J. Linn. Soc. 94: 127-142.

Rudall, P. 1987. Laticifers in Euphorbiaceae - a conspectus. Bot. J. Linn. Soc. 94: 143-163.

Rudall, P. 1994. Laticifers in Crotonoideae (Euphorbiaceae): homology and evolution. Ann. Missouri Bot. Gard. 81: 270-282.

Sherff, E.E. 1940. The concept of the genus. Bull. Torrey Bot. Club 67: 375-380.

Swofford, D.L. 1999. PAUP*: Phylogenetic Analysis Using Parsimony, version 4.0d65. Computer program. Sinauer,

Sunderland, Massachusetts.

Tokuoka, T. and H. Tobe. 1995. Embryology and systematics of Euphorbiaceae sens. Lat.: a review and perspective. J. Plant Res. 108: 97-106.

Venkata Rao, C. 1971. Anatomy of the inflorescence of some Euphorbiaceae. Bot. Not. 124: 39-64.

Webster, G.L. 1975. Conspectus of a new classification of the Euphorbiaceae. Taxon 24: 593-601.

Webster, G.L. 1994a. Classification of the Euphorbiaceae. Ann. Missouri Bot. Gard. 81: 3-32.

Webster, G.L. 1994b. Synopsis of the genera and suprageneric taxa of Euphorbiaceae. Ann. Missouri Bot. Gard. 81: 33-144.

Webster, G.L. and F.A. Rupert. 1973. Phylogenetic significance of pollen nuclear number in the Euphorbiaceae. Evolution 27: 524-531.

Webster, G.L. and B.D. Webster. 1972. The morphology and relationships of Dalechampia scandens (Euphorbiaceae). Amer. J. Bot. 59: 573-586.

Wurdack, K.J. and M.W. Chase. 1999. Spurges split: molecular systematics and changing concepts of Euphorbiaceae. In anonymous (eds.), Proceedings of the XVI International Botanical Congress, St. Louis, August 1-7, 1999, Abstracts, USA, pp. 142.