Asteropeia and Physena (CaryophyHales): A case study in comparative wood anatomy SHERWIN CARLQUIST Carlquist, S. (Santa Barbara Botanic Garden. 1212 Mission Canyon Road. Santa Barbara. California 93105. U.S.A.; e-mail: [email protected]). Asteropeia and Physena (Caryophyllales): A case study in comparative wood anatomy. Brittonia 58: 301-313. 2006.—Previous analyses of Asteropeia and Physena have not compared the wood anatomy of these genera to those of Caryophyllales 5.1. Molecular evidence shows that the two genera form a clade that is a sister group of the core Caryophyl­ lales. Synapomorphies of the Asteropeia—Physena clade inelude small circular alter­ nate pits on vessels, presence of vasicentric tracheids plus fiber-tracheids. presence of abaxial-confluent plus diffuse axial parenchyma, and presence of predominantly uniseriate rays. These features are analyzed with respect to habit and ecology of the two genera. Solitary vessels, present in both genera, are related to the presenee of va- sieentric tracheids. Autapomorphies in the two genera seem related to adaptations by Physena as a shrub of moderately dry habitats (e.g., narrower vessel elements, abun­ dant vasicentric tracheids. square to erect cells in rays) as compared to alternate char­ acter expressions that seem related to the arboreal habit and humid forest ecology of Asteropeia. The functional significance of vasicentric tracheids and fiber-tracheids in dicotyledons is briefly reviewed in the light of wood anatomy of the two genera. Key words: Asteropeiaceae. Caryophyllales. ecological wood anatomy, liber- tracheids, functional wood anatomy. Physenaceae, vasicentric tracheids. Two genera that comprise monogeneric by cladistic analysis of DNA data. Often families endemic to Madagascar, Asteropeia mentioned as examples are the tile cells (10 species according to International Plant (Chattaway, 1933) in rays of certain families Names Index) and Physena (two species), are of Malvales (now grouped by some as Mal­ of interest with respect to wood anatomy be­ vaceae s.l.) and the occurrence of vestured cause studies on woods of these genera pits in all families of Myrtales (Bailey, 1933; (Miller, 1975; Dickison & Miller, 1993) did Jensen et a!., 1998). Others are featured in not result in discernment of relationships, de­ Chapter 10 of Carlquist (2001). Assemblage spite thorough work. During much of the of wood data in the compilations by Sol- 20th century, wood anatomy has been consid­ ereder (1906) and by Metcalfe and Chalk ered as a potential source of systematically (1950) are undoubtedly valuable. How such and phylogenetically significant data. This compilations can be used, however, poses a view was inspired by Solereder's (1885) the­ question: if one is attempting to find the rela­ sis, "Uber den systematischen Wert der tionships of a genus or family, can one draw Holzstruktur," the title of which clearly advo­ on such data? If one does, one can choose cates this application. This idea pervades Sol- any number of taxonomic groups for compar­ ereder (1906), Metcalfe and Chalk (1950), ison. One must choose a reasonable number, and the several volumes of the second edition and so in the decades prior to molecular phy- of Anatomy of the Dicotyledons that have ap­ logenies, comparative wood anatomists relied peared in recent decades. Indeed, one can on natural systems as a guide to which com­ cite a number of fascinating correlations be­ parisons were most productive. Systems at­ tween wood anatomical data and the latest tempting to represent evolutionary relation­ incarnations of the natural system as revealed ships were many, however, and did not agree Brittonia. 58(4). 2006. pp. 301-313. ISSUED: 28 December 2006 © 2006. by The Now York Botanical Garden Press, Bronx. NY 10458-5126 U.S.A. 302 BRITJONIA [VOL. 58 with each other. Thus, comparative wood wood features likely represent symple- anatomists were induced both to choose more siomorphies within Caryophyllales s.l., and numerous groups and to rely on one particu­ which features are probably synapomorphies lar phylogenetic system more than on others. for the Asteropeia—Physena clade, and In attempting to elucidate the relationships of which features may be autapomorphies for Physena, Dickison and Miller (1993) com­ Asteropeiaceae and for Physenaceae, respec­ pared wood anatomy and other anatomical tively. Comparative wood anatomy cannot be features of that genus to those of Barbey- treated as a source of patterns or character aceae, Bonnetiaceae, Capparaceae, Clusi- states to be inserted into phylogenetic sys­ aceae, Eucommiaceae, Fagaceae, Flacour- tems. The biological nature of wood fea­ tiaceae, Passifloraceae, Sapindaceae, tures—evolutionary changes in wood anat­ Staphyleaceae, Theaceae, Urticaceae, and— omy related to changes in ecology and as a separate genus—Asteropeia. At that habit—is too important to be neglected by time, Asteropeia was thought to be thealean. wood anatomists. For example, many years so none of the comparisons were to ago, study of woods of helenioid Asteraceae Caryophyllales s.s. ("core Caryophyllales" essentially showed a uniform basic pattern, today). Nor were the families that were the variations of which all relate to ecology added to Caryophyllales, such as Droser- and habit (Carlquist, 1959). That conclusion aceae, Frankeniaceae, Nepenthaceae, Polygo- was extended to the entire family Asteraceae naceae, Tamaricaceae, on the basis of DNA- (Carlquist, 1966). Although in evolution of based phylogenies (such as those of Soltis et wood in dicotyledons as a whole the adaptive al., 2000), included in the comparisons of value of some features may be neutral, or Dickison & Miller (1993). The problem was may have been superceded by other changes, not in the ingenuity of the plant anatomists, one is best served by assuming that wood is a but in the limitations of pre-molecular phylo­ functioning structure, not an accretion of genetic systems. markers of historical events (Carlquist, With the advent of phylogenies of di­ 1975). Viewing wood features as a syndrome cotyledons based on sequencing of one or of tools for adaptation and survival may be more genes and using cladistic analysis, the difficult because one never has all of the in­ robustness of proposed relationships was formation one needs for such an enterprise, enormously heightened. In earlier decades, but in science, asking questions seems prefer­ the role of comparative wood anatomy was able to avoiding them. The materials used for thought to be not merely for detecting the re­ this study are limited by the collections avail­ lationships of poorly understood genera and able. One could wish that more material families, but also for offering data useful to might become available, but collection of construction of phylogenetic systems. How­ woods in Madagascar is extraordinarily diffi­ ever, data from wood anatomy at best cannot cult for logistical reasons and because of for­ offer the robust interpretations that DNA- est destruction. Collection of wood samples based phylogenies potentially offer, so the throughout the world, once a well-funded role of wood anatomy as a prime phyloge­ goal of forestry schools and institutes, has netic tool has faded. One can study data from now declined. Nevertheless, new examination wood anatomy in the light of a DNA-based of the Asteropeia and Physena specimens system, however, to learn and demonstrate available with both scanning electron mi­ how wood evolves. In the case of Asteropeia croscopy (SEM) and light microscopy does and Physena, Morton et al. (1997) found that yield reliable conclusions on the qualitative the two genera form a clade that is a member level, although conclusions at the quantitative of Caryophyllales s.l. and is the outgroup to level must be considered tentative. This is "core Caryophyllales" (Caryophyllales s.s.). true for any study based on xylarium samples The status of the two genera as comprising (as by far the majority are), because the monogeneric families in Caryophyllales has source of the sample (inside or outside of been accepted (APG, 2003). The present ex­ trunk; branch, etc.) is not specified in xylar­ amination of wood anatomy of the two gen­ ium wood samples. The only studies in wood era has, as goals, the demonstration of which anatomy that can validly claim quantitative 2006] CARLQUIST: ASTEROPEIA AND PHYSENA WOOD ANATOMY 303 TABLE I QUANTITATIVE WOOD FEATURES OF ASTEROPEIA AND PHYSENA. Species VG VD VM VL IL UK F\ ME .4. micraster 1.14 70 25 507 746 118 1.47 1420 A. multiflora 1.00 7X 15 487 703 106 1.44 2532 A. rhopaloides 1.00 1()4 8 647 769 145 1.19 8739 P. madagascariensis MADw-44404 1.00 41 24 387 686 238 1.77 661 P. madagascaru rm is MADw-4663 1 1.00 66 13 358 S74 166 2.44 441 Key to columns: VG. mean number of vessels per group: VD. mean vessel lumen diameter, m: VM. mean number of vessels per mm2 VL. mean vessel clement length, m; TL. mean imperforate tracheary element length, m: UR. mean uniseriate ray height, m: FV. F/V ratio (imperforate tracheary element length divided by vessel element length): ME. Mesomorphy Ratio (Vessel diameter times vessel element length divided by number of vessels per mm2). Collection data is given in Materials and Methods. significance are the extremely few that are linear row; pits can be multiseriate on any controlled for sampling and provenance; that given facet. No standard method for observa­ of Sastrapadja & Lamoureux (1969) can be tion and measurement of this feature exists. cited. Identification of vasicentric tracheids re­ quires observations of longisections com­ Materials and Methods bined with study of transactions; pit abun­ dance can be determined in imperforate All woods were available as dried samples. tracheary elements in both types of sections The collections are as follows: Asteropeia by careful observation of the frequency of micraster Hall., Analamazaotra Forest, 100 bordered pits. Although the wood data of km E of Antananarivo, Thouvenot 149 Miller (1975) and of Dickison & Miller (CTFw-362); A.