On the phylogenetic position of the New Caledonian endemic families Paracryphiaceae, Oncothecaceae, and Strasburgeriaceae: a comparison of molecules and morphology.
New Caledonia is one of Earth's great biodiversity hotspots (sensu Myers et al., 2000), primarily because of its high level of botanical endemism and its reputation as a refugium for some of the most ancient lineages of extant seed plants. Grande-Terre, the main island of the French territory, is only 390 km long by 50km wide (ca. 16,890 k[m.sup.2] in total area) but supports no fewer than 763 genera and 3,137 species of seed plants, of which 79% are endemic (Lowry, 1998). New Caledonia is isolated from all major landmasses, being located ca. 1,200 km east of Australia, 1,500 km northwest of New Zealand, and 1,800 km southeast of New Guinea. It is one of only a few Pacific islands that are not of volcanic origin and is thought to have separated from Gondwana during the early Cretaceous, approximately 65 million years ago. Its isolated location ca. 20 south of the equator has kept the island buffered from glaciation and extreme climate change, although controversial evidence suggests that all or part of the is land may have been submerged during the Tertiary (Paris et al., 1979). The vegetation, too, may have been kept isolated from the influence of exotic species invasions due to its occurrence on and adaptation to soils rich in toxic, heavy metals (Thorne, 1965; Lowry, 1998). In an essay on the island's floristic affinities, Thorne stated that "there is probably no region of comparable area in the world with such a rich, archaic, peculiar, and endemic seed-plant flora as that of New Caledonia" (1965: 2). Thorne's research and that of others have inspired plant morphologists around the world, including the late William C. Dickison, to make the pilgrimage to this small Pacific island in order to see for themselves its unique vegetation.
Shortly after beginning his tenure as a professor of botany at the University of North Carolina at Chapel Hill, Bill Dickison traveled to New Caledonia (in 1979 and 1981) for collecting purposes (Fig. 1). Those expeditions had an important impact on his career, for he subsequently published several dozen papers on the morphology and anatomy of New Caledonian plants, including members of Dilleniaceae and Cunoniaceae. However, Dickison was especially fascinated by the endemic angiosperm families from New Caledonia that he collected. Students of his "Plant Anatomy" course at UNC, Chapel Hill (including myself) recall Dr. Dickison's numerous references to these taxa in his lectures on the evolutionary trends and specialization of vessel elements in flowering plants.
New Caledonia is home to at least five endemic families of flowering plants: Amborellaceae (1 genus, 1 species), Phellinaceae (1 genus, 10 species), Paracryphiaceae (1 genus, 1 species), Oncothecaceae (1 genus, 2 species), and Strasburgeriaceae (1 genus, 1 species). One of these, Amborellaceae (represented by Amborella trichopoda Baillon) has been identified recently as the sister taxon to all remaining angiosperms and has consequently become the focus of numerous investigations into its natural history and fundamental biology. Although the other endemic families of New Caledonia may not be as famous as Amborellaceae, Bill Dickison recognized them as being equally fascinating, enigmatic, and poorly understood. From 1976 to 1986 he and his co-authors published a series of papers on the morphology and anatomy of Paracryphiaceae, Oncothecaceae, and Strasburgeriaceae. His research on these families attempted to provide objective data that would assist in clarifying the phylogenetic position of each, and his resul ts would serve as the primary basis for their placements in the comprehensive classification systems of Cronquist, Thorne, Takhtajan, and others.
The purpose of this article is to provide a review of the historical classification of and alternative phylogenetic relationships proposed for Paracryphiaceae, Oncothecaceae, and Strasburgeriaceae. These alternative systematic hypotheses, including those advocated by Dickison based on morphological and anatomical characters, will be compared and contrasted with hypotheses of phylogenetic relationships inferred by more recent cladistic analyses of molecular sequence data.
III. Paracryphiaceae Airy Shaw
Paracryphiaceae are a monotypic family represented solely by Paracryphia alticola (Schltr.) Steen. This species grows as a small tree up to 18 m tall at elevations of 600-1500 mon the highest mountains of New Caledonia (e.g., Mount Humboldt, Mount Panie, and Mount Ignambi). The plant produces inconspicuous, bisexual or staminate, sessile flowers with undifferentiated perianth along a branched, terminal, pubescent inflorescence (Fig. 1). The capsular fruits of Paracryphia are unique, consisting of 8-15 carpels that eventually separate at the base but remain attached to the central column distally. Although its reproductive morphology is highly specialized, its wood anatomy is considered one of the most primitive known in angiosperms. This combination of advanced and unspecialized features, as well as its geographical isolation, has kept the systematic affinities of this enigmatic family in question.
A. TAXONOMIC HISTORY
Paracryphia alticola has been treated in detail by Jeremie (1996) for the Flora of New Caledonia. It was originally described as Ascarina alticola in Chloranthaceae (Schlechter, 1907), but Baker (1921) placed it in its own genus, Paracryphia, and classified it within Eucryphiaceae. Although it does possess vessel elements, these are characterized by a number of primitive features, such as partially intact perforation plates (Carlquist, 1992). For this reason, several authors, including Gilg (1925), Bausch (1938), and Guillaumin (1948) considered P. alticola to be related to the vesselless families Winteraceae or Trochodendraceae. Swamy (1953) dismissed this notion but gave no alternative suggestion for its affinity. In 1965 Airy Shaw erected a new family, Paracryphiaceae, to accommodate the enigmatic taxon and placed it near Trochodendraceae. He even considered that it might best be classified within that family, but he also commented on the similarity in floral structure to another enigmatic taxon, Medusagyn e. Both Takhtajan (1973) and Cronquist (1981) recognized Paracryphiaceae as a distinct family. The former positioned it near Eucryphiaceae as others had previously, whereas the latter moved it to a placement within Theales and near Medusagynaceae, based on the detailed morphological investigation by Dickison and Baas (1977) discussed below. Most recently, Takhtajan (1997) elevated the family to the rank of monotypic order (Paracryphiales) but listed it next to Theales in Thorne's superorder Theanae. He stated that among angiosperms, "Paracryphia is one of the most remarkable ancient relicts" (1997: 160).
B. MORPHOLOGICAL/ANATOMICAL EVIDENCE
Although Swamy (1953) and others had published earlier studies of Paracryphia's floral and vegetative morphology, including details of its wood anatomy, Dickison and Baas (1977) recognized that the taxon was still poorly researched. They provided a comprehensive survey of the plant's morphology, anatomy, and potential systematic affinities. In several instances they demonstrated and corrected erroneous interpretations of the taxon in previous studies. Using a rigorous, systematic approach, they concluded that an affinity between Paracryphiaceae and any family in Magnoliidae (i.e., Chloranthaceae) is extremely weak. Likewise, based on a number of conflicting characters, they found very little support for the notion that the family might be related to Trochodendraceae, Eucryphiaceae, or Medusagyne. After discounting these historical hypotheses, their attention was focused on five other families that exhibit strong morphological similarities with Paracryphia: Theaceae, Actinidiaceae, Ericaceae, Aquifoliaceae, an d Sphenostemonaceae. A comparison of 17 characters was presented. They showed that Sphenostemon and Paracryphia share 15 characters and differ in only two: Sphenostemon has an indehiscent fruit and lacks the small, retirugate, tricolporate pollen of Paracryphia. In comparison, Paracryphiaceae were shown to share ten characters with Theaceae (but only six of them present in most members of the family), 13 characters with Actinidiaceae (seven of them present in most members), 11 characters with Ericaceae (seven of them present in most members), and 14 characters with Aquifoliaceae (seven of them present in most members). Although it would seem that Dickison and Baas (1977) had found the definitive evidence they were looking for to determine the exact sister family of Paracryphiaceae (i.e., Sphenostemonaceae), they hesitated in stating such. Instead, they safely concluded that Paracryphiaceae should continue to be treated as a distinct family and that "the families Actinidiaceae, Theaceae, and Sphenostemonaceae appear to us to be the closest extant relatives of Paracryphia ... These relationships should present an interesting challenge to contemporary system makers" (1977: 432).
C. MOLECULAR EVIDENCE
Figure 2 shows a minimum of six alternative positions historically suggested for the phylogenetic placement of Paracryphiaceae among angiosperms. The phylogeny was published by the Angiosperm Phylogeny Group (APG) (1998), and is based on a combination of molecular sequence data. It represents the most robust and conservative estimate of flowering plant relationships to date and resulted in a new ordinal classification being published by the APG (1998) for angiosperms. Within that system, Paracryphiaceae was treated as one of 25 families considered of "uncertain position" (APG, 1998) because molecular data for it were not available at that time. Likewise, Paracryphia was not included in the three-gene study of angiosperms published by Soltis et al. (2000) using rbcL, atpB, and 18S sequences. More recently, however, Savolainen et al. (2000) published their results of an analysis of rbcL sequences only, but for nearly all families of eudicots. Paracryphia alticola was included in this study, and it was shown to be sister to Sphenostemonaceae on a strongly supported branch within Dipsacales, as depicted in Figure 2.
IV. Oncothecaceae Kobuski ex Airy Shaw
This New Caledonian endemic family was considered to be represented by only one species, Oncotheca balansae Baill., until a second species, O. macrocarpa Guill., was described in 1982. The name of the latter was corrected later to 0. humbolitiana (Guill.) Morat. The trees grow to a height of 30 m and are found along river banks or other mesic areas, mostly at the extreme southern tip of New Caledonia's South Province. Ocotheca's small flowers are characterized by five distinct sepals, a sympetalous corolla of five petals, five epipetalous stamens alternate to the petals, and five carpels that form a superior, compound ovary with five distinct styles. The fruit is fleshy, indehiscent, and drupelike, with five locules and persistent sepals, as shown in Fig. 1. Oncotheca balancsae and 0. humboltiana differ from each other primarily in fruit shape, leaf size, and stamen morphology.
A. TAXONOMLC HISTORY
A comprehensive review of the systematic literature for Oncothecaceae was presented by Carpenter and Dickison (1976), and the family has been treated thoroughly by Morat and Veillon (1988). When it was originally described by Baillon (1891), Oncotheca was treated as a genus of Aquifoliaceae because of its resemblance to Phelline and Sphenostemon, which were classified within that family prior to being elevated to familial rank (Phellinaceae and Sphenostemonaceae). Later, Baillon (1892) suggested that Oncotheca might be related to either Cliftonia (now Cyrillaceae) or the family Ebenaceae.
For nearly half a century, botanists continued to recognize Oncotheca as a genus of either Aquifoliaceae or Ebenaceae, even though it was considered an isolated taxon by all. Guillaumin (1938, 1948) studied its morphology in detail and concluded that it would be treated best as Oncotheceae, a monotypic tribe of Ebenaceae. The family Oncothecaceae was eventually erected by Airy Shaw (1965), based on personal correspondence with C. Kobuski, who suggested that the taxon might be allied to Celastraceae or more likely to Theaceae. Airy Shaw (1965), himself, wrote that he favored placing Oncothecaceae in a position near Aquifoliaceae and listed a number of features, such as campanulate corolla and pentalacunar nodes in Oncotheca, that contradicted a placement near Theaceae. Takhtajan (1969, 1997) placed Oncothecaceae in Theales, as did Cronquist (1981), most likely in reaction to a thorough examination of morphology and anatomy published by Carpenter and Dickison (1976). In his most recent classification of angiosp erms, Takhtajan (1997) kept Oncothecaceae in Theales, suggesting an affinity not only to Theaceae, in general, but also to subfamily Ternstroemioideae, more specifically.
B. MORPHOLOGICAL/ANATOMICAL EVIDENCE
Baas (1975) and others before him had already documented the relatively unspecialized wood anatomy and vegetative morphology of Oncotheca before Carpenter and Dickison (1976) published their more comprehensive study of the genus. However, this latter study took into account both vegetative and reproductive morphology in assessing the relationships of the family. The authors concluded that poor soils rather than xeric habitat might explain Oncotheca's specialized coriaceous leaves with thickened cuticle, thick hypodermis, and compact spongy mesophyll. For this reason, emphasis on foliar anatomy in providing clues to the taxon's placement among angiosperms was downplayed and reasoned to be heavily influenced by environment rather than a result of phylogeny. They also commented on the advanced floral morphology of the genus, pointing out a number of features, such as stamens reduced to five in a single whorl and pendulous ovules reduced to two per locule, shared by Oncotheca and members of Theaceae subfamily Ter nstroemioideae. On the other hand, Carpenter and Dickison (1976) admitted that whereas both Oncothecaceae and Theaceae are characterized by relatively unspecialized wood anatomy (e.g., solitary vessels with small diameter, very oblique endwalls, and scalariform perforation plates in Oncotheca), the presence of multilacunar nodes in Oncotheca is in sharp contrast to the unilacunar nodes so characteristic of Theaceae. Nevertheless, they concluded that because a close relationship between Oncotheca and either Aquifoliaceae or Ebenaceae could not be justified, only Theaceae remained as the most likely candidate for closest living relative of Oncothecaceae.
With the discovery of O. mocrocarpa (an illegitimate name, later corrected to O. humboldtiana), Dickison (1982) again addressed the contributing role of morphology in the systematics of Oncothecaceae. He documented a number of characteristics that supported recognition of the new species from O. balansae. More importantly, this study provided Dickison with additional proof, in the form of astrosclerids found in the leaves of the new species, that Oncothecaceae might be related to Theaceae.
A few years later, Dickison (1986) would be provided with liquid-preserved flowering material of both species of Oncotheca. Dried herbarium specimens had provided most of the tissue used in his previous investigations. Dickison stated that, in his opinion, the floral anatomy and especially the pollen of Oncotheca were still in keeping with a Thealean relationship. In his conclusion, however, Dickison modestly admitted that he continued to be perplexed by Oncotheca's pentalacunar nodes and questioned whether Oncothecaceae might be related to Paracryphiaceae, Actinidiaceae, or Sphenostemonaceae (keep in mind that his 1977 study on Paracryphiaceae was completed a year after his first study on Oncothecaceae). He concluded his 1986 study by stating that "a more precise placement of the taxon must await accumulation of data from other fields of inquiry" (p. 258). The first molecular studies of angiosperm phylogeny were starting to yield results around this time, and Dickison must have felt that these new techniques held potential for settling the debate surrounding Oncotheca.
C. MOLECULAR EVIDENCE
Figure 3 shows at least four alternative phylogenetic placements historically suggested for Oncothecaceae, as well as the placement supported by molecular data. The family's exact position continues to be ambiguous, but the APG (1998) classified it as a member of Garryales along with Eucommiaceae, Garryaceae, and Aucubaceae, based on rbcL sequences (Savolainen et al., 2000). However, in their combined analysis of rbcL, atpB, and 18S, Soltis et al. (2000) found weak support for Oncotheca as sister to the entire euasterid I dade of Gentianales, Lamiales, and Solanales (with Garryales subsequently sister to all of them). Although the exact position of Oncothecaceae among angiosperms is still somewhat unclear, the family shows no affinity to Theaceae, Ebenaceae, Aquifoliaceae, or any other family suggested previously. Its phylogenetic position in euasterid I near Garryales represents a completely new hypothesis that will need to be tested with further molecular and morphological inquiry.
V. Strasburgeriaceae Engler & Gilg
Strasburgeria robusta (Viell. ex Panch. & Seb.) Guillaumin, the sole taxon in Strasburgeriaceae, is a large tree endemic to New Caledonia's high-elevation, montane rain forests. It is characterized by large, coriaceous, denticulate leaves and solitary flowers with ca. 10 spirally arranged sepals, five petals, 10 stamens, and ca. five connate carpels (Fig 1). The fruit of Strasburgeria is large, indehiscent, and fibrous, with persistent sepals and style and a distinctive odor of ripe apples.
A. TAXONOMIC HISTORY
Originally described as a species of Montrouziera (an endemic genus of Clusiaceae from New Caledonia), M. robusta was transferred into its own genus, Strasburgeria, by Baillon in 1876. Baillon's name for the species, S. calliantha, was corrected later to S. robusta by Guillaumin in 1942. In his description, Baillon suggested that the taxon could be related to Brexia (Saxifragaceae S.l. or Brexiaceae), Sapotaceae, or Theaceae. Other opinions followed, including those of Szyszylowicz (1893), Engler (1897) and Schlechter (1906), who suggested possible affinities to Erythroxylaceae, Ochnaceae, and Saxifragacene, respectively. A proposal to elevate the genus to familial rank was made by van Tieghem (1903), who placed Strasburgeria within a broadly defined Geraniales. The family Strasburgeriaceae was eventually established in 1925 by Engler to accommodate the enigmatic taxon. He positioned it next to Ochnaceae. Since that time, most systematists have chosen to classify Strasburgeria robusta either as a tribe of Och naceae (e.g., Beauvisage, 1920; Cronquist, 1981) or a distinct family closely related to Ochnaceae (e.g., Engler, 1925; Takhtajan, 1997).
B. MORPHOLOGICAL/ANATOMICAL EVIDENCE
The most thorough study of both the vegetative and reproductive anatomy of Strasburgeria robusta was completed by Dickison (1981) in order to help shed light on the taxonomic affinities of this isolated taxon. Dickison noted that various descriptions of the plant within the literature were incomplete or erroneous, and he hoped to remedy this problem with FAA-preserved material that had been secured just prior to his study. Detailed descriptions of every aspect of the plant's morphology and anatomy were presented, leading Dickison (1981) to conclude that "when direct comparisons of the organs and tissues of Strasburgeria, Ochnaceae, and Theaceae are made, there seems to be no doubt that a true phylogenetic relationship exists between these taxa" (1981: 577). Although this would seem to be a definitive "case-closed" statement by the author, Dickison nevertheless commented on several discrepancies between the morphology of Strasburgeriaceae and these other families. For example, he admitted that no species of Oc hnaceae have stamens or fruits similar to those of Strasburgeria. Neither are there Ochnaceous or Theaceous taxa with such primitive wood anatomy, specialized pollen morphology, or randomly organized higher order leaf venation as Strasburgeria. Moreover, he pointed Out that the ovules of Ochnaceae are tenuinucellate, whereas those of Strasburgeria are crassinucellate. These rather striking differences were substantial enough for Dickison to concur that the taxon warranted familial rank separate from Ochnaceae, but he concluded that it was similar enough, overall, to be kept in a position adjacent to that family, "in the current widely accepted manner" (1981: 579).
C. MOLECULAR EVIDENCE
As depicted in Figure 4, the phylogenetic position of Strasburgeriaceae can be mapped to at least five different locations on the APG (1998) cladogram, based on historically hypothesized, alternative positions. In their ordinal classification of angiosperms, however, the APG (1998) listed Strasburgeriaceae as one of its 25 families of unknown placement because of insufficient molecular data. Complete sequences of Strasburgeria were also unavailable to Soltis et al. (2000) for their combined rbcL, atpB, and 18S phylogeny of flowering plants. Savolainen et al. (2000), however, did include an rbcL sequence for S. robusta in their phylogenetic reconstruction of eudicots. That study found Strasburgeria to be closely related to Ixerba brexioides A. Cunn. (Ixerbaceae) within or sister to Crossosomatales, which in turn is sister to Geraniales. The authors stated that "if these relationships are confirmed (especially using additional data for Geissolomna and Strasburgeria), then Crossosomatales could be expanded to in clude Geissolomataceae, Ixerbaceae, Strasburgeriaceae, and perhaps Aphloiaceae" (Savolainen et al., 2000: 272).
To further test this result, both the plastid gene atpB and the nuclear ribosomal large sub-unit (185) were sequenced from Strasburgeria robusta. Genomic DNA was extracted from silica gel-dried leaf tissue collected in New Caledonia (K. Cameron 9839, NY), and the target loci were amplified and sequenced at the New York Botanical Garden. The primers and general methods employed for gene amplification and automated sequencing have been described by Soltis et al. (2000). The two new sequences (GenBank #AF502596 and #AF502597), together with the previously published rbcL sequence of Strasburgeria (GenBank #AJ403007), were added to a seven-taxon subsample of the Soltis et al. (2000) three-gene matrix to represent Crossosomatales and Geranium (Geraniales) as the outgroup. A branch and bound search was executed under the parsimony criterion using PAUP* 4.0b10 (Swofford, 2002) to find all most parsimonious trees. In addition, a total of 1,000 bootstrap (bts) replicates were performed to quantify the level of support for the recovered relationships. A single tree was recovered and is presented in Figure 5. The result of this new three-gene analysis corroborates the sister relationship between Strasburgeria and Ixerba with 100% bts support. Furthermore, the pair is sister (87% bts) to a second dade supported by 100% bts that contains Staphyleaceae, Stachyuraceae, and Crossosomataceae. The latter two genera are sister to each other with 99% bts, as shown in Figure 5. Savolainen et al. (2000) showed weak evidence from rbcL sequences that Geissolomataceae may also be part of this larger dade, but it was not included in the three-gene study here because of insufficient data.
D. COMPARISON OF ANATOMICAL AND MOLECULAR EVIDENCE
The molecular evidence overwhelmingly indicates that Strasburgeriaceae are sister to another monotypic family, Ixerbaceae. Ixerba brexioides is endemic to the low, montane forests in the far north of New Zealand's North Island, where it is known commonly by its Maori name, "tawari." Honey made from bees that pollinate tawari flowers has a distinctive flavor and is marketed as a specialty item throughout New Zealand (Salmon, 1980). Like Strasburgeria, Ixerba also has a confused taxonomic history. In addition to being treated as an enigmatic, monotypic family (Ixerbaceae), it has been treated most commonly within either Escalloniaceae or together with Brexia in Brexiaceae. Classification of these families and their constituent genera has been one of the most troublesome problems for systematists (for an overview of the controversies, see J. Reveal's online discussion of Grossulariaceae at <http://life.umd.edu/emeritus/reveal/PBIO/pb450/rosi01.html#gros>). No direct affinity between Ixerba and Strasburgeriaceae has ever been suggested, although it is interesting to recall that when Baillon (1876) originally named Strasburgeria, he suggested that it might be related to Brexia, an idea that was never considered further.
Fortunately, the wood anatomy of Ixerba brexioides has been documented in detail by Patel (1973), and this study clearly shows anatomical features supporting the sister relationship between Strasburgeria and Ixerba. As seen in Table I, the two taxa have strikingly similar wood that, for the most part, differs only in the length and width of vessel elements. Moreover, there are similarities in flower and fruit morphology between the two taxa, but also enough differences to separate the two genera into different families. For example, both taxa have persistent sepals, clawed petals, and persistent style with punctiform stigma (see Fig. 1), yet Ixerba has a coriaceous, dehiscent fruit, whereas that of Strasburgeria is fibrous and indehiscent.
Finally, it is worth noting that fossil pollen from Tertiary deposits in southern Australia, Tasmania, New Zealand, and New Caledonia is directly comparable to modern Strasburgeria robusta pollen (Jarzen & Pocknall, 1993), although nearly half its size. The distribution and age of these fossils, known as Bluffopollis scabratus (Couper) Pocknall & Mildenhall, indicate that the common ancestor of Strasburgeria and Ixerba most likely evolved prior to the breakup of Gondwana and that it was widely distributed on that supercontinent.
When comparing the phylogenetic relationships of taxa based on results of cladistic analysis of molecular data with historical hypotheses of relationships based on morphology, at least three different scenarios may be encountered. In the first scenario, molecular data may be found to confirm the "traditional" phylogenetic position of a taxon or settle the debate between two alternative placements. This has been shown to be the case for Paracryphiaceae, a taxon which has had a convoluted taxonomic history but for which the molecular data agree with the most recent morphological data (published mostly by Dickison & Baas, 1977) in positioning the family next to Sphenostemonaceae.
A second scenario is exemplified by Oncothecaceae. In this case, the molecular data refute all possible placements for a taxon and indicate a completely new phylogenetic position never considered previously. Oncothecaceae may be sister to Garryales or to the entire euasterid I clade (sensu Soltis et al., 2000), based on recent cladistic analyses of DNA sequences. But in either case the family appears not to be allied to Aquifoliaceae, Ebenaceae, Theaceae, or any other family previously hypothesized. These are the results that excite plant systematists (and sometimes the media) most. They are also the results that lend themselves to further investigation by the process of "reciprocal illumination." Although not obvious at the present time, characters undoubtedly will be discovered or simply recognized for the first time to support this newly proposed relationship. Dickison (e.g., 1982, 1986) commented more than once that Oncotheca's pentalacunar nodes just did not fit with its placement in Theales. Perhaps thi s would be a good character to examine first in searching for synapomorphies with Garryales.
Finally, a third possible scenario may play out when comparing molecular with morphological data. A taxon such as Strasburgeriaceae may find itself positioned in an unexpected clade on a molecular phylogram, thereby refuting its "accepted" classification. However, if the historical literature is examined carefully, one often discovers that this very same relationship was predicted long ago but was long since forgotten. Although Dickison's (1981) morphological and anatomical studies of Strasburgeriaceae were meticulous and insightful, he was nevertheless forced to compare the taxon primarily with Ochnaceae, because this had been the overwhelmingly agreed upon sister family cited by most systematists before him. As discussed above, Dickison clearly had reservations about classifying Strasburgeria within or near Ochnaceae, but he never considered Brexiaceae (as was originally suggested by Baillon in 1876) as an alternative. If he had, he most likely would have considered Brexia first (and would have correctly di smissed it as a close relative), but he might also have dug a little deeper and stumbled upon another segregate of that family, Ixerba. This is the most likely sister to Strasburgeria, based on newly presented molecular data. In many ways, this scenario is the most ironic, but also most satisfying, one for a systematist. It allows the objective analysis of DNA molecules to vindicate those botanists from the last century who had little more technology at their disposal than a light microscope, hand lens, or well-trained eye. Bill Dickison would have appreciated this irony as much as anyone.
Table I A comparison of wood anatomy characters for Ixerba and Strasburgeria Ixerba brexioides Strasburgeria robusta Growth rings indistinct Growth rings absent Vessels solitary, sometimes in 2s Vessels solitary, sometimes in 2s or rarely 3s or rarely 3s Vessels angular in outline Vessels angular in outline Vessel elements average 39.7 mm Vessel elements average 95.4 mm wide, 1090 mm long wide, 2249 mm long Endwalls nearly vertical, Endwalls nearly vertical, exclusively scalariform exclusively scalariform Perforation plates average 37 bars Perforation plates average 27 bars per plate per plate Intervascular pits opposite Intervascular pits opposite Parenchyma diffuse, scanty Parenchyma diffuse, apotracheal and paratracheal scanty paratrachial Rays heterogeneous, 1-3 cells wide Rays heterogeneous, 1-3 cells wide Crystals absent Crystals absent
This paper was delivered in Albuquerque, New Mexico at a Botany 2001 annual meeting symposium honoring the late William C. Dickison, who served as one of my doctoral co-advisors. First and foremost, I wish to thank him for his wisdom, guidance, inspiration, and friendship during my years as a graduate student at the University of North Carolina at Chapel Hill and afterwards. I also must acknowledge his wife, Marlene Dickison, who provided photographs and stories of Bill's field trips to New Caledonia. In addition, the following individuals were of special assistance in the preparation of this article: Tanguay Jaffre of IRD, Noumea, New Caledonia; Ken Wurdack of the New York Botanical Garden; and Susan Whitfield and Bill Burk of the University of North Carolina at Chapel Hill. The photographs of Paracryphia and Strasburgeria were taken by P. Endress and P. Lowry, respectively. This study was funded by the New York Botanical Garden and the Lewis B. and Dorothy Cullman Foundation,
VIII. Literature Cited
Airy Shaw, H. K. 1965. Diagnoses of new families, new names, etc. for the seventh edition of Willis's Dictionary. Kew Bull. 18: 249-273.
APG (Angiosperm Phylogeny Group). 1998. An ordinal classification for the families of flowering plants. Ann. Missouri Bot. Gard. 85: 531-553.
Baas, P. 1975. Vegetative anatomy and the affinities of Aquifoliaceae, Sphenostemon, Phelline, and Oncotheca, Blumea 22: 311-407.
Baillon, H. 1876. Stirpes exoticae novae. Adansonia 11: 366-373.
-----. 1891. Sur le nouveau genre Oncotheca. Bull. Mens. Soc. Linn. Paris 2: 931-932.
-----. 1892. Histoire des plantes. Vol. 11. Hachette, Paris.
Baker, E.G. 1921. Systematic account of the plants collected in New Caledonia and the Isle of Pines by Prof. R. H. Compton, M.A., in 1914, I. Dicotyledons. Polypetalac. J. Lion Soc., Bot. 45: 264-325.
Bausch, J. 1938. A revision of the Eurcruphiaceae. Kew Bull. 1938: 317-349.
Beauvisage, L. 1920. Contribution a l'etude anatomique de la famille des Ternstromiacees. Unpublished thesis.
Carlquist, S. 1992. Pit membrane remnants in perforation plates of primitive dicotyledons and their significance. Amer. J. Bot. 79: 660-672.
Carpenter, C. S. & W. C. Dickison. 1976, The morphology and relationships of Oncotheca balansae. Bot, Gazette 137: 141-153.
Cronquist, A. 1981. An integrated system of classification of flowering plants. Columbia Univ. Press, New York.
Dickison, W. C. 1981. Contributions to the morphology and anatomy of Sirasburgeria and a discussion of the taxonomic position of the Strasburgeriaceae. Brittonia 33: 564-580.
-----. 1982. Vegetative anatomy of Oncotheca macrocarpa, a newly described species of Oncotheeaeeae. Bull. Mus. Natl. Hist. Nat., B, Adansonia 4: 177-181.
-----.1986. Further observations on the floral anatomy and pollen morphology of Oncotheca (Oncothecaccac). Brittonia 38: 249-259.
-----.& P. Baas. 1977. The morphology and relationships of Paracryphia (Paracryphiaceac). Blumea 23: 417-438.
Engler, A. 1897. Nur zweifelhaft zu den Ochnaceac gehorige Gattung. P. 245 in A. Engler & K. Prantl (eds.), Die naturlichen pflanzenfamilicn. Nachtrage zum II-IV. W. Engelmann, Leipzig.
-----.1925. Strasburgeriaceae. Pp. 87-89 in A. Engler & K. Prantl (eds.), Die naturlichen pflanzenfamilien, Vol. 21. W. Engelmann, Leipzig.
Gilg, E. 1925. Eucryphiaceac. Pp. 47-50 in A. Engler & K. Pranti (eds.),. Die naturlichen pflanzenfamilien, Vol. 21. W. Engelmann, Leipzig.
Guillaumin, A. 1938. Observations morphologiqucs et anatomiques sur le genre Oncotheca. Rev. Gen. Bot. 50: 629-635.
-----.1942. Contribution a la fore de la Nouvelle-Caledonie, 77. Plantes des collecteurs divers. Bull. Mus. Hist. (Paris), Ser. 2, 14: 144-150.
-----. 1948. Flore analytique et synoptique de Ia Nouvelle-Caledonie: Phanerogames. Office de Ia recherche scientifique coloniale, Paris.
Jarzen, D. M. & D. T. Pocknall. 1993. Tertiary Bluffopollis scabratus (Couper) Pocknall & Mildenhall, 1984 and modem Strasburgeria pollen: A botanical comparison. New Zealand J. Bot. 31:185-192.
Jeremie, J. 1996. Paracryphiaceae. Fl. Nouvelle-Caledonie 20: 107-111.
Lowry, p. p., II. 1998. Diversity, endemism, and extinction in the flora of New Caledonia: A review. Pp. 18 1-206 in C.-I. Peng & P. P. Lowry II (eds.), Rare, threatened, and endangered floras of the Pacific Rim. Institute of Botany, Academia Sinica, Taipei, Taiwan.
Morat, P. & J.-M. Veillon. 1988. Oncothecaceae. Fl. Nouvelle-Caledonie 15: 90-98.
Myers, N., R. A. Mittermeier, C. C. Mittermcier, C. A. B. da Fonseca & J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403 (6772): 853-858.
Paris, j. p., p. Andreieff & J. Coudray. 1979. Sur Iage. Eocene superieur de Ia mise en place de Ia nappe ophiolitique de Nouvelle-Caledonie. Compt. Rend. Acad. Sci. Paris, Ser. B, 288: 1659-166 1.
Patel, R N. 1973. Wood anatomy of the dicotyledons indigenous to New Zealand, II. Escalloniaceac. New Zealand J. Bot.11: 421-434.
Salmon, J. T. 1980. The native trees of New Zealand. A. H. & A. W. Reed, Wellington, New Zealand.
Savolainen, V., M. F. Fay, D. C. Aihach, A. Backlund, NI. van der Bank, K. M. Cameron, S. A. Johnson, M. D. Lledo, J.-C. Pintaud, NI. Powell, M. C. Sheanan, D. E. Soltis, P. S. Soltis, P. Weston, W. M. Whitten, K. J. Wurdack, Chase. 2000. Phylogeny of the eudicots: A nearly complete familial analysis based on rbcl gene sequences. Kew Bull. 55: 257-309.
Schlechter, R. 1906. Beitrage zur kenntnis der flora von Neu-Kaledonien. Bot. Jahrb. Syst. 39: 114.
-----. 1907. Beitrage zur kenntnis der flora von Neu-Kaledonien. Bot. Jahrb. Syst. 39: 1-274.
Soltis, D. E., P. S. Soltis, M. W. Chase, NI. E. Mart, D. C. Albach, NI. Zanis, V. Savolanine W. H.
Hahn, S. B. Hoot, M. F. Fay, M. Axtell, S. NI. Swensen, L. M. J. Kress, K. C. Nixon & J. S. Farris. 2000. Angiosperm phylogeny inferred from 18S rDNA rbcL, and atpB sequences. Bot. J. Linn. Soc. 133: 381-461.
Swamy, B. G. L. 1953. Comments on Ascarina alticola Schlechter. Proc. NatI. Inst. Sci. India 19: 143-147.
Swofford, D. L. 2002. PAUP* 4.OblO: Phylogenetic analysis using parsimony. Sinauer Associates, Sunderland, MA.
Szyszylowicz, I. 1893. Theaceae (Temstroemiaceae). Pp. 175-192 in A. Engler & K. Prantl (eds.), Die natfirlichen pflanzenfamilien 3(6). W. Engelmann, Leipzig.
Takhtajan, A. 1969. Flowering plants: Origin and dispersal. Smithsonian Institution Press, Washington, DC. 1973.
-----. 1973. Evolution und Ausbreitung de Bliitenpflanzen. G. Fischer Verlag, Stuttgart, Germany.
-----. 1997. Diversity and classification of flowering plants. Columbia Univ. Press, New York.
Thorne, R. 1965. Floristic relationships of New Caledonia. State Univ. Iowa Stud. Nat. Hist. 7:1-14.
Van Tieghem, P. 1903. Sur Ic genre Sirasburgdria considere comme type d'une famille nouvelle, les Strasburgeriacees. J. Bot. (Morot) 17: 198-204.
|Printer friendly Cite/link Email Feedback|
|Author:||Cameron, Kenneth M.|
|Publication:||The Botanical Review|
|Date:||Oct 1, 2002|
|Previous Article:||Eulogy for Bill Dickison.|
|Next Article:||Comparative stem anatomy in the subfamily Cactoideae.|