IAWA Bulletin n.s., Vol. 13 (I), 1992: 3-16

WOOD ANATOMY AND STEM OF CHLORANTHUS; SUMMARY OF WOOD ANATOMY OF CHLORANTHACEAE, WITH COMMENTS ON RELATIONSHIPS, VESSELLESSNESS, AND THE ORIGIN OF MONOCOTYLEDONS

by

Sherwin Carlquist Rancho Santa Ana Botanic Garden and Department ofBiology, Pomona College, Claremont,Califomia 91711 , U.S.A.

Summary In contrast to the monopodial Ascarina and dicotyledons. While these cases are theoreti­ Hedyosmwn, Ch/oranthus and Sarcandra are cally possible, the histological and ecological sympodial. Sarcandra and Coerectus have seenarios that must be hypothesised for these woody canes of finite duration, whereas other events are ignored by cladists; most of these species of Ch/oranthus have shoots of one seenarios are unlikely for reasons explored year's duration ; these latter species have sec­ here, although a few are still worthy of con­ ond year wood only on rhizome s, not on up­ sideration. Stern endodermis is reported for right shoots. Rhizome portions transitional to three species of Ch/oranthus . upright sterns were selected for study. Chlo­ Key words: Ascarina, Chloranthaceae, Chlo­ ranthus erectus has abundant septate fibre­ ranthus , endodermis in sterns, Hedyos­ tracheids, C. jap onicus none, and two other mum, monocotyledon origins, Piperales, species a few. Ch/oranthus (and Sar candrai Sarcandra , vessel evolution, wood anat­ have rays of two distinct sizes in wood: rays omy . that are extensions of primary rays, and uni­ seriate and biseriate rays in fascicular areas. Introduction Wood anatomy of each of the four genera can be characterised, and is summarised in the The present study offers new data on wood form of a key. Except for primitiveness of anatomy of the genus Ch/oranthus. This pa­ vessels, wood of Chloranthaceae is very sim­ per is the fourth in a series that deals with ilar to that of Lactoridaceae and Piperaceae, wood of the family ; the other papers have and this probably indicates a close phyletic dealt with Sar candra (Carlquist 1987), As­ relationship. The large rays of chlorantha­ carina (Carlquist 1990), and Hedyosmwn ceous wood, little modified from prim ary (Carlquist 1992a). A summary of wood anat­ rays and with upright cells predominantly, are omy of the family is offered; other papers that indicative of some degree of herbaceousness have contributed to this summary are those of and some degree of secondary woodiness. Thierry (1912), Swamy & Bailey (1950), and Scattered bundles and multilacunar node s, Swamy (1953). characteristics of monocotyledons, are absent The inherent phylogenetic interest of CWo­ in Chloranthaceae but present in Piperaceae. ranthaceae is very great. The farnily has ethe­ The sympodial habit ofCh/oranthus and Sar­ real oil cells as well as other feature s that candra, and the presence of vessel s in roots place it unmistakably in Magnoliales sensu but not in sterns of Sarcandra are conditions lato (now superorder Magnoliidae of some like those basic to origin of monocotyledons. authors). However, the location of the farnily The possibility that Chloranthaceae are close within the order is not at all certain: several to Piperales and that these groups are close to recent treatment s consider it in a separate or­ origin of monocotyledons should be consid­ der (Chloranthaies) of the superorder Magno­ ered. Some cladists have hypothesised that liidae (e.g. , Lammers er al, 1986), which secondary vessellessness is polyphyletic in merely states the uncertainty of its relation-

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 4 IAWA Bulletin n.s., Vol. 13 (1), 1992 ships. Endress (1987), in a review of fea­ Whether vessellessness is primaryor sec­ tures of the family, cites more numerous ondary in particular dicotyledon phylads is an resemblances of Chloranthaceae to Hamame­ issue that has received comments in recent lidales, which lack ethereal oll cells and are years. For several reasons, Chloranthaceae are not placed in Magnoliidae, than to Laurales of pertinent in this regard, and a section of this Magnoliidae. Chloranthaceae have some high­ paper is devoted to a discussion of this matter. ly primitive features (e.g., wood) plus some Stern endodermis is reported here for three highly specialised features (floral structure). species ofChloranthus. This feature is SC8ICe Few will agree with Leroy's (1983) idea that in angiosperms, so occurrences are worthy of the inflorescences of Hedyosmum are actual­ being recorded. ly multistaminate primitive flowers (see the response to that idea by Endress 1987). The Materials and Methods phylogenetic problems posed by Chlorantha­ Stems and roots of ChJoranthus serratus ceae are not elose to solution at present, (Carlquist 15684 RSA) were collected in the The habit, wood, and vascular bundles of forest of Mt. Unzen, Kyushu, Japan, in 1982 Chloranthus in particular, and Chloranthaceae and preserved in formalin-acetic-alcohol. The in general, contain features relevant to con­ kindness of Prof. Mikio Ono in aiding my sideration with respect to origin of mono­ field work is gratefully acknowledged. Mate ­ cotyledons. Although Chloranthaceae have rial of other species of Chloranthus was de­ not been considered elose to origin of mono­ rived from herbarium specimens: C. erectus cotyledons because of such features as their (Stone 12116 KLU), Templer Park, Selangor, rather specialised flowers, opposite leaves, Malaysia; C. japonicus (Gorovoy 8 June and cylindrical arrangement of bundles, the 1968 RSA),N of the River Komarovka, family is definitely worthy of consideration in Primorye Terr., Siberia, USSR;C. multi­ this respect. Features ancestral to origin of stachys (Sino-American Botanical Expedition monocotyledons include sympodial growth 1185 RSA). habit, short-lived upright stems, minimal cam­ Stems of Chloranthus erectus were boiled bial activity, wide primary rays , and vessels in water, stored in 50% aqueous ethyl alco­ in roots only (tracheids only in other organs): hol, and sectioned on a sliding microtome. these are features of Sarcandra (and, except for Sections for study by SEM were dried be­ distribution of vessels, Ch/oranthus). Trim­ tween clean glass slides. Material of C. ser­ ery, basic to flowers of monocotyledons, is ratus was infiltrated, embedded in paraffin evident in flowers of Chloranthaceae (female and sectioned. Material of C. japonicus and flowers of Hedyosmum have three tepals ). C. multistachys was from herbarium speci­ Chloranthaceae have a range of habits: mens and was treated with 2% NaOH to ex­ Ascarina and Hedyosmum are monopodial pand tissues, then stored in 50% aqueous shrubs to trees, whereas Chloranthus and Sar­ ethyl alcohol, infiltrated, embedded in paraf­ candra have a sympodial growth form; Sar­ fin, and sectioned. Sections of all species candra, C. erectus, and C. spicatus have up­ were stained with a safranin-fast green com­ right shoots that last for several years and can bination. Paraffin seetions for SEM study branch somewhat, whereas the remaining were mounted on aluminum stubs followed species of Chloranthus form upright stems by removal of paraffin by means of xylene. that last only a single year. This range of habits Macerations of all species were prepared by is obviously of interest with respect to wood means of Jeffrey's fluid and stained with anatomy, An attempt is made here to charac­ safranin. terise the wood of the species and genera stud­ Terms used are according to the IAWA ied. In this regard, the reader should be aware Committee on Nomenclature (1964). Primary that species names commonly encountered in wall presence in perforations is referred to as Chloranthus and Sarcandra have been sup­ membraneremnants (Carlquist 1992b). Vessel planted in accord with the studies ofVerdcourt diameter is taken as lumen diameter at widest (1985). A synonymy is offered in the Materi­ point. In scanning a section with the light mi­ als and Methods section. croscope to determine vessels per mm -, ray

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 5 areas were not excluded (but only secondary nonstoried. Starch present in ray cells and xylem was included). occasionally in septate fibres. The species names in Chloranthus and Chloranthus japonicus (Figs. 12-15). Sarcandra are according to Verdcourt (1984, Wood diffuse-porous (Fig. 12), although the 1985): Chloranihus erectus (Buch.-Ham .) second year's wood (on rhizomes and bases Verdcourt (= C. ofjicinalis Blume); C.japo­ of upright stems) has smaller vessels, often nicus Siebold; C. multistachys Pei; C. serra­ irregular in shape and size (Fig. 15, left third tus Roem & Schult; Sarcandra chloranthoi­ of photograph), whereas the first year's ves­ des Gardner (= S. irvingbaileyi Swamy); S. sels are arranged in orderly rows (Fig. 12). glabra (Thunb.) Nakai [= C. brachystachys Vessels at the end of first year wood are nar­ Blume, S. hainanensis (Pei) Swamy & Bai­ rower and more tracheid-like (Fig. 14, mid­ ley]. Sarcandra thus consists of two species; dle); some may actually be tracheids. During Chloranthus, which has not been monograph­ the second year's growth (Fig. 14, top), wider ed as a genus, contains about 10. vessels are intermixed with narrower vessels; narrower vessels bear scalariform end wall pitting typical of wider vessels (Fig. 15). Anatomical descriptions of wood Vessels as seen in transection in large groups Chloranthus erectus(Figs. l-ll). Diffuse­ by virtue of absence of imperforate tracheary porous (Fig. 1). Vessels mostly solitary (Fig. elements. Mean vessel diameter, 33 um . 1); mean number ofvessels per group, 1.15. Mean vessel wall thickness, 2.2 um. Mean Mean diameter of vessels, 32 lUD. Mean num­ vessel element length, 1188 um. Perforation ber of vessels per mm-, 107. Mean vessel plates all scalariform (Fig. 20). Mean number wall thickness, 2.0 lUD. Mean vessel element of bars per plate, 76. Lateral wall pitting of length, 1207 um, Perforation plates scalari­ vessels transitional, opposite, or alternate. form (Fig. 6), mean number of bars, 78. Axial parenchyma not present in the ordinary Membrane remnants in perforations varying sense; parenchyma cells between vessels in from extensive (Fig. 7) to nearly absent (Fig. metaxylem grade into rays within the fasci­ 8). Remnants present as aseries of flakes cular areas. Rays of two sorts, the large rays (Figs. 9, 10) or as bands in the lateral ends of essentially extensions of primary rays (Fig. the perforations (Fig. 11). Lateral wall pitting 13, left and right), whereas uniseriate or bi­ of vessels consisting of circular pits about 5 seriate rays are present in fascicular areas um in diameter, these pits usually in uniseri­ (Fig. 13, centre) . Mean height of the large ate vertical rows, corresponding to vessel rays more that 5 mm; mean height of uni­ walls facing fibre-tracheids. All imperforate seriate rays, 1265 um, Ray cells all upright tracheary elements are septate fibre-tracheids (Fig. 15, bottom). Ray cell walls lignified but (Figs. 4, 5), pits sparse with apertures thin (1.2 lUD thick). Wood nonstoried. Starch obliquely oriented (Figs. 4, 5). Pits bordered, present in ray cells. pit cavity diameter varying from 2 um to 5 Chloranthus multistachys (Fig. 19).Wood um, Mean fibre-tracheid diameter, 21 um, diffuse-porous, although the second year's Mean fibre-tracheid length, 1333 lUD. Axial wood has vessels irregular in shape and size. parenchyma not observed. Rays formed from Second year wood present only in rhizoma­ interfascicularareas wide (Fig. 2), mean width tous sterns and at bases ofupright stems. Ves­ at widest point, 11.4 cells. Mean ray height sels often not solitary, often in radial rows, more than 5 mm, accurate estimation on the mean number of vessels per group, 1.93. basis of sections impossible. Occasional uni­ Mean vessel diameter, 39 um, Mean number seriateand biseriaterays presentin axial xylem of vessels per mm", 64. Mean vessel wall formed from fascicular areas; mean height bi­ thickness, 1.8 um, Mean vessel element seriate rays, 1004 um; mean height uniseriate length, 1248 J.1m. Perforation plates scalari­ rays, 700 lUD. All ray cells upright, prominent­ form, membrane remnants common in perfo­ ly elongate (Fig. 3). Ray cell walls lignified, rations. Mean number of bars per perforation wall thickness about 3 lUD. Pits interconnect­ plates, 72. Lateral wall pitting of vessels op­ ing ray cells conspicuously bordered. Wood posite or alternate. Imperforate tracheary ele-

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 6 IAWA Bulletin n.s., Vol. 13 (1),1992

11 'I

Figs. 1-5. Wood seetions of Chloranthus serratus. - 1: Transeetion, showing abundanee of fibre-traeheids eompared to vessels . - 2: Tangential seetion; two large multiseriate rays are evident. - 3: Radial seetion, showing upright ray eells and, at right, septate fibre-traeheids. - 4: Two septate fibre-traeheids from radial seetion; pit sizes are different. - 5:SEM photograph of septate fibre-tracheids from radial seetion; pit apertures are oblique. - Figs. 1 & 2, seale above Fig. 1 (divisions = 10 11m); Fig. 3, seale above Fig. 3 (divisions = 10 11m); Fig. 4, seale above Fig. 4 (divisions = 10 11m); Fig. 5, bar at upper right = 5 11m.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 7

Figs. 6-11. SEM photographs ofperforations plates from radial sections of Chloranthus erectus wood. - 6: Entire perforation plate; remnants can be seen in perforations. - 7: Perforation plate with maximal retention of membrane remnants. - 8: Perforation plate with only a few thread-like remnants of membrane. - 9: Flake-like membrane remnants in perforations. - 10: Small vestiges of membranes along margins of perforations.- 11:Membrane remnants in lateral ends of per­ forations . - Fig. 6. bar =5 11m; Figs. 7-11, bar in upper right of Fig. 7 = 111m.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 8 IAWA Bulletin n.s ., Vol. 13 (1),1992

Figs. 12-15. Sections of wood of Chloranthus japonicus. - 12: Transection, pith at right; two years accumulation of secondary xylem are present. - 13: Tangential section; large multiseriate rays at left and right, uniseriate rays in fascicular portion, centre. - 14: Transection; cells with thicker walls, center, represent end of first year's growth, second year's wood above. - 15: Radial section; the irregular cells in left third of photographs represent the second year's wood. - Fig. 12 & 13, scale above Fig. 1; Fig. 14 & 15, Scale above Fig. 3.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 9

Figs. 16-20. Wood sections ofChloranthus. - 16-18: C. serratus. -16: Transection; a single year's wood is shown. - 17: Radial section; vessels with scalarifonn perforation plates, above; ray cells, below center; septate fibre-tracheids, bottom. - 18: Perforation plate from tangential section; membranes in perforations are intact. -19: C. multistachys, SEM photograph ofperfo­ rations showing membrane remnants. - 20. C. japonicus; SEM photograph ofperforations near­ ly free of membrane remnants . - Fig. 16, scale above Fig. 1; Fig. 17, scale above Fig. 3; Fig. 18, bar = 5 11m. Figs. 19, scale in Fig. 8. Fig. 20, scale in Fig. 7.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 10 IAWA Bulletin n.s. , Vol. 13 (1), 1992 ments are an septate fibre- traeheids, less abun­ Casparian strip which stains rose with safran­ dant in number than vessel elements, with pits in in contrast to the remainder of the cell wall, sparser than on lateral walls of vessels, pits which stains green with fast green. An endo­ about 4 um in diameter. Mean fibre-tracheid dermis of this sort has not previously been length, 1352 um, Axial parenchyma not ob­ reported in stems ofChloranthaceae, although served. Larger mu1tiseriate rays are essenti­ Thierry (1912) figures an endodermis for C. ally extensions of primary rays; mean height oldhamii Solms. The endodermis he figures of multiseriate rays more than 5 mm. Mean is like a layer of thin-walled selereids rather width of multiseriate rays, 6.2 cells. A few than cells with thin primary walls bearing uniseriate or biseriate rays are present within Casparian strips. the axial xylem formed in the fascicular areas. Stern endodermis is not a common pheno­ All ray cells are upright. Ray cell walls are menon in dicotyledons, but the instances re­ lignified, about 1.4 um thick . Pits among ray ported thus far do not appear to have phylo­ cells simple or very slightly bordered. Wood genetic significance within dicotyledons at nonstoried. Stareh present in ray cells. large (presence and absence within a single Chloranihus serratus (Figs. 16-18). Wood family may be significant). For example, diffuse-porous, although diminution in vessel stern endodermis with clear Casparian strips diameter is evident at the end of the growth occurs in helianthoid Asteraceae such as ring (representing a single year) (Fig. 16). Fitchia (Carlquist 1957). Metcalfe and Chalk Vessels often in radial rows (Figs. 16, 17). (1950) report a number of genera of Lamia­ Mean number of vessels per group, 2.2. Mean ceae with various types of endodennis, in­ vessel diameter, 44 um. Mean number ofves­ eluding some with typical Casparian strips. sels per mm-, 193. Mean vessel wall thick­ Asteraceae and Lamiaceae are not considered ness , 2.7 um, Mean vessel element length, families extremely elose to each other, and 1264 um, Perforation plates scalarifonn (Fig. stern endodennis has been reported, in any 17, above). Many perforations observed to case , in only a small number of dicotyledon contain membrane remnants that are intact farnilies . The occurrence of stem endodermis (Fig. 18). Mean number ofbars per plate, 82. in monocotyledons may be more significant Lateral wall pitting mostly alternate, rarely with relation to the stern endodermis in Chlo­ opposite. Imperforate tracheary elements all ranthus. Stern endodermis may be found in septate fibre-tracheids (Fig. 17, bottom); pit rhizomes of Araceae such as Acorus. The re­ diameter 2 um, borders vestigial. Mean fibre­ ports of stern endodermis in Chloranthus are tracheid diameter, 25 um; mean length, 1372 from bases of upright sterns ; these regions um, Axial parenchyma not observed. Multi ­ are very elose to the prostrate stern portions seriate rays wide, little altered extensions of that form the rhizomes, and are histologically primary rays; mean height of multiseriate rays essentially the same as the rhizomes. more than 5 mm. Uniseriate and biseriate rays present in axial secondary xylem fonned in Differences in wood of ChJoranthus fascicular areas. Most ray cells upright, a few The wood of the fcnr species of Chloran­ procumbent cells in which horizontal dimen­ thus differs with respect to quantitative fea­ sion only slightly exceeds vertical dimension tures, but these are of relatively little sig­ present (Fig. 17). Ray cell walls lignified nificance. The qualitative differences are (Figs . 16, 17). Ray cell walls about 2.0 um distinctive and relate to habit. The selection of thick. Pits among ray cells simple or with nar­ the species was designed to cover the range row borders. Wood nonstoried. Starch present within the genus in woodiness. Chloranthus in ray cells. erectus is the woodiest, with canes that add wood along their length, although that wood Stem endodermis accumulation is finite : only 2-3 mm thick­ The layer of cells external to phloem fibres ness of secondary xylem. Chloranthus japo­ is differentiated as an endodermis in sterns of nicus, C. multistachys, and C. serratus have C. erectus, C. multistachys, and C. serratus. upright sterns that cannot be called woody: The endodermis in these species has a typical they do not develop secondary xylem along

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 11 most of their length after the first year, and and stature of the upright sterns. Chloranthus only the rhizomes and the transitions between japonicusrepresents the ultimate end point in rhizomes and upright sterns add secondary xy­ such aseries. lem after the first year, and this added xylem The series formed by these Chloranthus tends to be irregular in configuration (e.g ., species can be read from woody to herbace­ Fig. 12). Of the three species mentioned, C. ous, although there is no evidence for uni­ japonicus qualifies as the most nearly herbace­ directional evolution of that sort. The impor­ ous (its range is also the most northerly). tant correlation is with the cane-like habit, The four species fall into three groups with which, in turn, is related to a sympodial habit. respect to fibre-tracheid presence: Chloran­ Fibre-tracheids would not likely be lost if a thus erectus: in the transition from metaxylem monopodial habit were retained. Aseries such to secondary xylem, parenchyma among ves­ as that cited above is instructive with relation sels is supplanted by fibre-tracheids; as in to origin of monocotyledons, because they most dicotyledon woods, imperforate trache­ have xylem lacking in dimorphism of trache­ ary elements (septate fibre-tracheids in this ary elements (strength provided instead by genus) are much more abundant than vessel extraxylary fibres) and sympodial growth elements. Chloranthus spicatus (Thunb.) habit: both of these conditions have been Maxim. agrees with C. erectus with respect completely achieved by C.japonicus. to fibre-tracheid morphology and presence (Takahashi 1985). - Chloranthus multi­ Chloranthaceae and the origin of MO­ stachys and C. serratus: in the transition from nocotyledons; relationships of Chloran­ metaxylem to secondary xylem, parenchyma thaceae among the vessels is supplanted by fibre-tra­ cheids, but the number of vessels increases at The preceding paragraph shows how two this point, so that fibre-tracheids are relatively interrelated conditions that must be hypothe­ few in number compared to the vessel ele­ sised as basic to origin of monocotyledons: ments, and the secondary xylem is composed sympodial growth habit and monomorphism mostly of vessel elements. - Chloranthus ja­ of tracheary elements. However, the mono­ ponicus: in the transition from metaxylem to morphism in C. japonicus is achieved by secondary xylem, parenchyma among vessels presence of vessels without any fibre-trache­ is supplanted by rays composed of upright ids, yet a primitive monocotyledon ought to cells. Fibre-tracheids are absent, and the axial have not vessels throughout, but tracheids portion of the secondary xylem consists throughout the plant body or in roots only. wholly of vessels. Cheadle's (1942) data shows that the most The shift from abundance to paucity to ab­ primitive monocotyledons already have ves­ sence of fibre-tracheids is a significant trend. sels in roots only, although Cheadle (1953) The most obvious correlation is with woody hypothesises independent origin of vessels in versus herbaceous. Chloranthus erectus is monocotyledons and dicotyledons. Tracheids clearly the woodiest, and tallness of its canes throughout aplant but with vessels in roots is and the fact that they can branch during a a condition found in Sarcandra (Carlquist second or third year relate to the greater me­ 1987). Ifa species of Sarcandra had the her­ chanical support provided by the abundant baceous habit of Chloranthus japonicus, it fibre-tracheids (C. spicatus is similar to C. would, in terms of bundle structure and tra­ erectus in habit and fibre-tracheid presence: cheary element type and distribution through­ Takahashi 1985). In the other species, which out the plant satisfy the criteria required for a are relatively herbaceous, fibre-tracheids are dicotyledon ancestral to a monocotyledon (the few or (C.japonicus) none. Addition of sec­ Sarcandra type of dicotyledon was not con­ ondary xylem in these species occurs only on sidered by Cheadle, 1953, in his hypothesis rhizomes and the bases of upright shoots trans­ of origin of vessels in monocotyledons). Cri­ itional to the rhizomes; for these species, teria other than these would have to be met phloem fibres likely offer sufficient mechani­ for plants like Chloranthaceae to be ancestral cal support for the relatively short duration to monocotyledons, however, and extant

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 12 IAWA Bulletin n.s., Vol. 13 (1),1992

Chloranthaceae do not satisfy all of these al activity, combined with wide primary rays, criteria. Arber (1917) claimed there was vestigial cam­ Among the criteria not satisfied by Chlo­ bial activity within vascular bundles of mono­ ranthaceae as potential ancestors of mono­ cotyledons, although this statement has been cotyledons are the opposite leaves, cylinder disputed by some (e.g., Katherine Esau , per­ of bundles, and unilacunar-two trace nodes sonal communication). Philipson et a/. (1971) (Swamy 1953). These nodes, however, are have carefully examined this question, and believed to be the most primitive type in di­ concluded that some monocotyledon genera cotyledons (Bailey 1956). The family Pipe­ iGloriosa, Veratrum) have cambial activity raceae, however, has alternate leaves and that adds vascular tissue within a bundle scattered bundles. These are significant as an during a second growing season. Philipson example of how the leaf position, nodal anato­ et al. (1971 ) note that monocotyledons lack my, and bundle configuration likely basic to both intrafascicular cambia and initiation of monocotyledons are present in a phylad close­ rays within bundles. However, the presence Iy related to Chloranthaceae provided that ofresidual cambia in Liliaceae (because ofits Piperaceae are, in fact, closely related to CWo­ primitive xylem and other reasons a family ranthaceae. Are Chloranthaceae and Pipera­ often considered as primitive within mono­ ceae closely related? cotyledons) is significant in that it indicates The wood anatomy of Lactoridaceae proves derivation of monocotyledons from a group to be essentially identical to that of Piperaceae of dicotyledons with ar least some cambial (Carlquist 1990b). This by itself proves no­ activity. Vascular bundles like those cited by thing about the relationships of Chlorantha­ Philipson et al. (1971) could be derived from ceae, but it does widen our concepts of Piper­ vascular bundles like those in Sarcandra or ales , in which order Lactoridaceae must be Chloranthus merely by reduction of cambial placed ifevidence from wood is as significant activity - a matter of degree rather than a as I believe it to be. The wood features of CWo­ presence or absence difference. ranthaceae are very much like those of Lacto­ The large multiseriate rays of Chlorantha­ ridaceae and Piperaceae: the on1y two differ­ ceae, Lactoridaceae, and Piperaceae are most­ ences of significance are the presence of ly little-altered continuations of primary rays, scalariform perforation plates and nonstoried Rays in all of these families are composed of wood in Chloranthaceae; perforation plates upright cells exclusively or predominantly. are simple (occasionally scalariform in Lacto­ Both features suggest paedomorphosis, using ridaceae) and wood is storied in Lactoridaceae the criteria of Carlquist (1962) . In fact, a spe­ and Piperaceae. These are both differences of cies of Piperaceae (Macropiper excelsum evolutionary level rather than systematic diver­ Miq.) was cited as having the paedomorphic gence per se. Both storied wood and simple descending curve of vessel element length perforation plates have evolved in many phy­ with age, a curve contrary to the curve for lads independently, so they are much more typical woody dicotyledons (Carlquist 1962). likely to be homoplasies than synapomor­ Paedomorphosis characterises wood of di­ phies. Such interesting features as presence cotyledons that are either secondarily woody of living septate fibre-tracheids with vestigi­ or are herbaceous or very nearly herbaceous ally bordered pits, wide tall multiseriate rays (Carlquist 1988a). Because the herbaceous little altered from primary rays, and upright habit is basic to monocotyledons, the occur­ ray cells exclusively or nearly so are common rence of paedomorphic rays and vessel ele­ to the three families. Certainly the wood anato­ ments in Chloranthaceae and possibly allied my of the three families is compatible with families offers yet another reason why Chlo­ the idea that they are more closely related to ranthaceae should be considered in connec­ each other than each is to the nearest out­ tion to origin of monocotyledons. Although group. there is no decisive evidence on whether the Lass of cambial activity is thought to be monopodial (Ascarina , Hedyosmum) or sym­ characteristic of origin of monocotyledons. podial (Chloranthus, Sarcandra) habit is prim­ Certainly Chloranihus shows minimal cambi- itive in Chloranthaceae, there are reasons to

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 13 believe the arboreal habit in Hedyosmum is the roots branch, are not to be expected. In derived (e.g., the axial parenchyma type in Sarcandra; there is a transitional kind of sym­ Hedyosmum, the most arboreal genus, is podial habit, in which rhizomes, roots, and more specialised than axial parenchyma con­ sterns last more than one year (although only ditions elsewhere in the family; male flowers for a few years at most). One might expect of Hedyosmum are quite reduced). that a vessel condition more like the monoco­ Other features of Chloranthaceae bear con­ tyledon condition than the dicotyledon condi­ sideration with respect to origin of monocot­ tion would be present in Sarcandra because yledons. Trimery, characteristic of flowers of of the brevity of secondary xylem activity in monocotyledons, is prominent in flowers of sterns and roots. In a sympodial dicotyledon Saururaceae of the Piperales (Tucker 1985) with Ionger lasting sterns and roots (e.g., a and in flowers of Lactoridaceae (Carlquist woody Piper), interconnection between ves­ 1964). The flowers of Chloranthaceae are sels in roots and those of sterns on which the highly reduced, but Chloranthus retains tri­ roots have formed is to be expected. mery in its androecium, and Hedyosmum Irone looks at recent attemptsto use cladis­ retains trimery in tepals of female flowers tics to elucidate relationship of families men­ (Endress 1987). Monosulcate pollen grains, tioned above as well as other families with the type basic to monocotyledons, occur in notably primitive features, one sees quite Ascarina(Swamy 1953); the other genera of disparate results. The cladograms offered by Chloranthaceae have types that are closely Young (1981) , Lammers et al. (1986) and related to monosulcate pollen grains. Burger Donoghue and Doyle (1989) dealing with (1977) hypothesised that Piperales are closely these families differ marked1y from each related to monocotyledons. The evidence ad­ other with respect to key families . For ex­ duced by Burger, including such features as ample, Lammers et al. (1986) show Piperales floral anatomy, is different from that cited by and Chloranthales as distant outgroups, iso­ the above authors, but it deserves recon­ lated from each other, of Magnoliales (in sideration. which they include Lactoridaceae). Donoghue The significance of the sympodial habit and Doyle (1989), on the other hand, show for distribution of vessels in monocotyledons families traditionally placed in Magnoliales as cornpared to dicotyledons has not been and/or Laurales branching off their clade sufficiently appreciated. Bailey (1944, 1953) both above and below Hamamelidales. Don­ believed that vessels originated in secondary oghue and Doyle (1989) place Chlorantha­ xylem of roots and sterns of dicotyledons, ceae close to Trimeniaceae, Austrobaileya­ whereas Cheadle (1942) presented evidence ceae, Monimiaceae, and Amborellaceae; they that in monocotyledons, vessels originated in place Lactoridaceae close to Piperaceae and the root and then progressed upward in the Lactoridaceae but also, curiously, close to plant phyletically. Because there is a continu­ Aristolochiaceae). One especially pertinent ity between the root and the stem secondary feature of the Donoghue & Doyle scheme is xylem formed in a given year in a typical that Piperaceae, Saururaceae, and Lactori­ rnonopodial dicotyledon, simultaneous vessel daceae (but not Chloranthaceae) are placed origin is likely: for efficient conduction, ves­ close together and close to origin of mono­ sels in the root should be intercontinuous with cotyledons, as represented by the single vessels formed in the stern. This is not true in monocotyledon family they include, Lilia­ a monocotyledon because of the sympodial ceae. It would be easy to say that Donoghue habit. With the sympodial habit of monoco­ and Doyle (1989) are 'correet' except that tyledons, new roots are constantly initiated they have misplaced Chloranthaceae (a close on the newly-formed rhizome portions, so relationship between Chloranthaceae and Tri­ that all roots are adventitous and no roots meniaceae, also suggested by Endress, 1987, have secondary xylem. Therefore , intercon­ is hypothesised by Donoghue and Doyle). nections between vessels of roots, not form­ The matter cannot be resolved by such state­ ed ontogenetically at the same time as the ments. The three cladograms of primitive di­ tracheary elements of the stems from which cotyledon families cited are strikingly differ

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 14 IAWA Bulletin n.s., Vol. 13 (1), 1992 ent, and there is no reason to believe that any The axial parenchyma and ray type of As­ of the workers has been less than diligent. carina are more primitive thanthe correspond­ The problem appears to lie with the relictual ing condition in Hedyosmum. Ascarina has nature of these families. In such families, nonseptate fibre-tracheids or even tracheids. characters of phyletic significance are rela­ Ascarina has more primitive inflorescences tively few - so many have been lost in the than Hedyosmum in the interpretation of course of evolution. The paucity of signifi­ Endress (1987). cant characters means that an attempt at cla­ The diffuse axial parenchyma of Sarcan­ distic analysis will by this circumstance over­ dra marks that genus as at least as primitive stress some characters. The solution to this as Ascarina. The presence of vessels only in problem lies not so much in attempts to reuse roots of Sarcandra and its diffuse parenchyma information (largely from gross morphology) are best interpreted as primitive in compar­ the information already at hand, but to con­ ison to Chloranthus, which has vessels in struct cladograrns with the aid of new, more both roots and stems, clear' differentiation highly significant,kinds of data. Various types between vessels and septate fibre-tracheids, of DNA data are likely to yield data of de­ and no appreciable amount ofaxial paren­ cisive phyletic importance. Therefore, my re­ chyma. I am dubious about the repon of commendation is that hypotheses such as the diffuse axial parenchyma in C. spicatus by idea that Chloranthaceae may be close to Pipe­ Takahashi (1985); such parenchyma does not rales, that Lactoridaceae is in Piperales, and appear in Takahashi's figures, and must be that Chloranthaceae and Piperales are close to quite scarce, if present origin of monocotyledons, should be com­ If one attempts to say whether the sym­ pared to results provided by new DNA data. podial genera Chloranthus and Sarcandra are more primitive than the monopodial genera Phylesis within Chlorantbaceae Ascarina and Hedyosmum on the basis of Because of divergences among species of wood anatomy, one cannot reach a conclusive Chloranthus with respect to wood anatomy, distinction , Bisexual flowers, as in Chloran ­ as discussed above, xylary features ofChlo­ thus and Sarcandra, are traditionally consid­ ranthus are more difficult to summarise than ered more primitive than unisexual flowers are those of the other genera. However, the (Ascarina, Hedyosmum), but that is a single salient differences can be summarised in the character that should not be stressed. As with form of a key: phylogeny of families potentially related to Chloranthaceae, DNA data that can be anal­ Rays of two sorts (wide, tall rnultiseriate rays are ysed so as to yield polarity in phyletic status extensions of prirnary rays, uniseriate and biseriate are very much needed. rays are formed by cambium of the fascicular areas); cambial action restricted because of syrnpodial habit; ray cells contain starch. Vessellessness in diootyledons Vessels in roots only, not in upright sterns; axial The presence of intact pit membrane rem­ parenchyma diffuse Sarcandra nants in perforations of vessel elements in Vessels in stems and roots; axial parenchyma ab- Chloranthus serratus is interesting in that it sent or nearly so Chlorantbus indicates how readily the normal pattem of Multiseriate rays of various widths, not just biseriate lysis of pit membranes ean be reversed. How­ and wide rnultiseriate; cambial action indefinite be­ ever, this oceurrence should not be read as a cause of rnonopodial habit; droplets or amorphous de­ phyletie loss of vessels. As the preparations posits of yellow-staining cornpounds in ray cells. very clearly show, the tracheary elements are Axial parenchyma diffuse (or in short tangential still dimorphie, septate fibre-traeheids are still bands) ; uniseriate rays present in a11 stems; fibre- present, and thus even if intact pit membranes tracheids rnostly nonseptate Ascarina oceurred in all perforations of all individuals Axial parenchyrna vasicentric; uniseriate rays pre­ of C. serratus, the most important criteria of sent only in young sterns; fibre-tracheids all septate vessellessness would not have been satisfied. Hedyosmum In order for seeondary vessellessness to be

Downloaded from Brill.com10/04/2021 03:29:42AM via free access Carlquist - Anatomy of Chloranthus 15 anained, not only would pit membranes have phaeaceae, we cannot decide which hypo­ to be retained uniformly in all perforations, thesis is correct. If sterns ofSarcandra have the imperforate tracheary elements accompa­ less secondary xylem than do the roots, then nying the vessel elements would have to possibly vessels may be absent because the vanish - a change unlikely as long as there is development of wood is foreshortened in selection for mechanical strength in stems sterns - although there are other possible sce­ (should both types of tracheary elements be narios, such as selective advantage of vessels retained, most criteria of vessellessness would in roots versus sterns. not be satisfied). Even if such a change did Some cladists, such as Young (1981) or occur, one would have to have retrograde Doyle and Donoghue (1989) have proposed evolution in order to erase any differences be­ events of secondary vessellessness in woody tween end wall pitting and lateral wall pitting dicotyledons. They evade the responsibility in the vessel elements - differences which of explaining the histological, ecological, and characterise the vast majority of vessels in habital seenarios that must be proposed to ac­ dicotyledons. Thus, secondary vesselless­ count for such proposed events ofvesselloss. ness in a woody dicotyledon is highly un­ The above discussion is intended to show what likely unless one invents a quite extreme some of these unappreciated complexities may scenario . be. The extremeness of the seenarios that must An extreme scenario that could lead to ves­ be proposed should be considered by those sellessness has been described in the case who feel that more parsimonious cladograms of two nearly vesselless species of Ephedra can be achieved by such hypotheses. (Carlquist 1988b). A phylad in which both vessel elements and tracheids were present Referenoes would have to adapt to an extremely dry cold habitat, like the Ephedra species, in which Arber, A. 1917. On the occurrence of intra­ tracheids are more advantageous than vessel fascicular cambium in monocotyledons. elements at restricting the embolisms created Ann. Bot. 31: 41-45. by drought or freezing. Then such a phylad Bai1ey, LW. 1944. The development of ves­ would have to shift back to more moderate se1sin angiosperms and its significance in conditions - in fact, all of the extant vessel­ morphological research. Amer. J. Bot. 31: less dicotyledons are restricted to the most 421-428. mesic conditions possible. Bailey, LW. 1953. Evolution of tracheary Another scenario for secondary vesselless­ tissue of land plants. Amer. J. Bot. 38: ness - but one that will apply in only a few 245-254. limited instances and that does not really Bailey, LW. 1956. Nodal anatomy in retro­ involve loss of vessels is suggested by the spect. J. Arnold Arb. 37: 269-287. sterns of Sarcandra or stems of Nymphaea­ Burger, W.C. 1977. The Piperales and the ceae. Comparative work on secondary xylem monocots. Alternate hypotheses for the of dicotyledons led Bailey (1944, 1953) to origin of monocotyledons. Bot. Rev. 43: conclude that in dicotyledons, vessels origi­ 345-393. nated in secondary xylem (simultaneously in Carlquist, S. 1957. The genus Fitchia (Com­ roots and sterns), then progressed into prima­ positae). Univ. California Pub!. Bot. 29: ry xylem. There are dicotyledons in which 1-144. vessels occur in secondary xylem but not in Carlquist, S. 1962. A theory of paedomor­ primary xylem. If such a dicotyledon lost phosis in dicotyledonous woods. Phyto­ secondary xylem entirely, it could become morphology 12: 30-45. vesselless. But the outcome of this scenario Carlquist, S. 1964. Morphology and relation ­ would be an herbaceous plant. Nymphaea­ ships of Lactoridaceae. Aliso 5: 421-435. ceae, which are vesselless and have no sec­ Carlquist, S. 1987. Presence of vessels in ondary xylem, might be an example of this wood ofSarcandra (Chloranthaceae); com­ scenario or they might be primitively ves­ ments on vessel origins in angiosperms. selless - until we know the ancestry of Nym- Amer.1. Bot. 74: 1765-1771.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access 16 IAWA Bulletin n.s., Vol. 13 (1),1992

Carlquist, S. 1988a. Comparative wood anat­ Lammers, T.G., T.F. Stuessy & M. Silva omy. Springer Verlag, Berlin, Heidelberg. O. 1986. Systematic relationships of the Carlquist, S. 1988b. Near-vessellessness in Lactoridaceae, an endemic family of the Ephedra and its significance. Amer. J. Juan Femandez Islands, Chile. Plant Syst. Bot. 75: 598-601. Evol. 152: 243-266. Carlquist, S. 1990a. Wood anatomy of As­ Leroy , J.-F. 1983. The origin ofangiosperms: carina (Chloranthaceae) and the tracheid­ an unrecognized ancestral dicotyledon, vessel element transition. Aliso 12: 667­ Hedyosmum, with a strobiloid flower is 684. living today . Taxon 32: 169-175. Carlquist, S. 1990b. Wood anatomy and rela­ Metcalfe, C.R. & L. Chalk. 1950. Anatomy tionships of Lactoridaceae. Amer. J. Bot. of the dicotyledons. Clarendon Press, Ox­ 77: 1498-1505. ford . Carlquist, S. 1992a. Wood anatomy ofHedy­ Philipson, W.R., J.M. Ward & B.G. Butter­ osmum (Chloranthaceae) and the tracheid­ field. 1971. The vascular cambium. Its de­ vessel element transition. Aliso (in press) . velopment and activity . Chapman & Hall, Carlquist, S. 1992b. Pit membrane remnants London. in perforation plates of primitive dicotyle­ Swamy, B.G.L. 1953. The morphology and dons and their significance. (Submitted to relationships of Chloranthaceae. 1. Amold Amer.J. Bot.) Arb. 34: 375-408. Cheadle, V. 1942. The occurrence and types Swamy, B.G.L. & LW. Bailey. 1950. Sar­ of vessels in the various organs of the candra, a vesselless genus of the Chloran­ plant in the Monocotyledoneae. Amer. J. thaceae.J, Amold Arb. 31: 119-129. Bot. 29: 441-450. Takahashi, A. 1985. Wood anatomical studies Cheadle, V. 1953. Independent origin of ves­ of Polycarpicae. L Magnoliales. Sei. Rep. sels in the monocotyledons and dicotyle­ Osaka Univ. 34: 29-83.

dons. Phytomorphology 3: 23-44. Thierry, R. 1912. Contribution ä l'etude ana­ Donoghue, M.J . & J.A. Doyle. 1989. Phylo­ tomique des Chloranthacees, Thesis, Uni­ genetic analysis of angiosperms and the versity of Paris. relationships of Hamamelidae. In: Evo­ Tucker, S. 1985. Initiation and development lution, systematics and fossil history of of inflorescence and flower in Anemopsis Hamamelidae, vol. 1. Introduction and califomica (Saururaceae). Amer. J. Bot. 'Iower' Hamamelidae (eds. P.R. Crane & 72: 20-31. S. Blackmore): 17-45. Clarendon Press , Verdcourt, B. 1984. The correct name for the Oxford. southem Indian and Sri Lankan species of Endress, P.K. 1987. The Chloranthaceae: Sarcandra Gardner (Chloranthaceae). Kew reproductive structures and phylogenetic Bull. 39: 66. position. Bot. Jahrb. Syst. 109: 153-226. Verdcourt, B. 1985. Notes on Malesian Chlo­ IAWA Committee on Nomenclature. 1964. ranthaceae. Kew Bull. 40: 213-224. Multilingual glossary of terms used in Young, D.A. 1981. Are the angiosperms wood anatomy. Verlagsbuchanstalt Kon­ primitively vesselless? Syst, Bot. 6: 313­ kordia, Winterthur, Switzerland. 330.

Downloaded from Brill.com10/04/2021 03:29:42AM via free access