IAWA Journal, Vol. 19 (4), 1998: 383—391

WOOD AND STEM ANATOMY OF PETIVERIA AND RIVINA (CARYOPHYLLALES); SYSTEMATIC IMPLICATIONS by Sherwin Cariquist

Santa Barbara Botanic Garden, 1212 Mission Canyon Road, Santa Barbara, CA 93105, U.S.A.

SUMMARY

Petiveria and Rivina have been placed by various authors close to each other within Phytolaccaceae; widely separated from each other but both within Phytolaccaceae; and within a segregate family (Rivinaceae) but still within the order Caryophyllales. Wood of these monotypic genera proves to be alike in salient qualitative and even quantitative features, including presence of a second cambium, vessel morphology and pit size, nonbordered perforation piates, vasicentric axial parenchyma type, fiber-tracheids with vestigially bordered pits and starch contents, nar row multiseriate rays plus a few uniseriate rays, ray cells predominant ly upright and with thin lignified walls and starch content, and presence of both large styloids and packets of coarse raphides in secondary phloem. Although further data are desirable, wood and stem data do not strongly support separation of Petiveria and Rivina from Phyto laccaceae. Quantitative wood features correspond to the short-lived perennial habit of both genera, and are indicative of a xeromorphic wood pattern. Key words: Caryophyllales, Centrospermae, ecological wood anatomy, Phytolaccaceae, successive cambia, systematic wood anatomy.

INTRODUCTION

Phytolaccaceae s.l. have been treated quite variously with respect to generic contents. Gyrostemonaceae, once included in Phytolaccaceae, contain glucosinolates and have other features that require their transfer to Capparales (Goldblatt et al. 1976; Tobe & Raven 1991; Rodman et al. 1994). Other familial segregates from Phytolaccaceae recognized by most recent authors include Achatocarpaceae, Agdestidaceae, Bar beuiaceae, and Stegnospermataceae (Heimerl 1934; Rodman et al. 1984; Takhtajan 1987; Thorne 1992); all of these families except Achatocarpaceae, however are known to contain betalains and are retained within Caryophyllales as satellite families of Phytolaccaceae. Behnke (1997) makes a case for recognition of Agdestidaceae, Aizoaceae, Barbeuiaceae, Nyctaginaceae, Phytolaccaceae, Rivinaceae, and Sarcobata ceae as a suborder, Phytolaccineae, of Caryophyllales. Molecular data are interpreted by Rettig et al. (1992), Manhard and Rettig (1994), Downie and Palmer (1994), and 384 IAWA Journal. Vol. 19 (4), 1998

Downie et a!. (1997) as showing that Nyctaginaceae and Phytolaccaceae either are cladistically unresolved or else that Phytolaccaceae are polyphyletic with Nycta ginaceae intervening between portions of Phytolaccaceae. Heimerl (1934) placed Petiveria and Rivina next to each other within Phytolac caceae. The cladogram of Brown and Varadarajan (1985) separated Petiveria and Rivina widely but included both within Phytolaccaceae. Hershkovitz (1991) said that Agdestis “possibly links Phytolaccaceae s. s. to Rivinaceae” (Petiveriaceae is gener ally regarded as a synonym of Rivinaceae). Careful examination of wood and stem anatomy of Phytolaccaceae s. 1. and other Caryophyllales becomes valuable in view of the unstable taxonomic history of the genera and families described above. The review of Gibson (1994) shows that wood is poorly known for most families of Caryophyllales. Among the notable exceptions are Didiereaceae (Rauh & Dittmar 1970) and Cactaceae (various publications by Gibson; see Gibson 1994). In order to develop data sets on wood and stem anatomy of Caryophyllales, monographs on wood and stem anatomy have been provided for Caryophyllaceae (Carlquist 1995), Portulacaceae and Hectorellaceae (Carlquist 1997); Basellaceae (Carlquist, in press); andAgdestis (Cariquist, in press). Study of outgroups of Caryophyllales is also desirable, and one such family, Plumbaginaceae, has been monographed (Cariquist & Boggs 1996). Petiveria and Rivina have been selected for the present study because of the difference among authors as to whether they might be closely related or not, and because they, like Agdestis, are elements of suborder Phytolaccineae, the portion of the order I am currently studying. Petiveria and Rivina are also of interest with respect to ecological wood anatomy; they are weedy short-lived perennials, and thus their root system is likely to experi ence greater fluctuation in water availability than those of tree or shrub species. Peti veria alliacea L. is native from Florida to southern Brazil, whereas Rivina humilis L. is native from Texas to Argentina and Chile; both genera have become weeds in other areas (Heimerl 1934). The nature of cambial activity in species with successive cambia, such as Phyto laccaceae s. 1., has been subject to varied terminology and interpretations (see Gibson 1994 and Carlquist, in press). Consequently, any genus with successive cambial ac tivity potentially offers valuable information towards resolution of these problems.

MATERIAL AND METHODS

Liquid-preserved stems and roots of Petiveria alliacea were provided from cultivated material in the Heidelberg Botanic Garden (no. 2317197). Stems of Rivina humilis (Carlquist 8162, SBG) were collected from plants growing in a weedy area of the University of Hawaii campus, Honolulu. Both of the above specimens were preserved in 50% aqueous ethanol. In addition, dried stems of Petiveria alliacea, collected in the Panama Canal Zone (MADw-35452), were provided by the Forest Products Labo ratory, Madison, Wisconsin. The liquid-preserved material of Petiveria was sectioned according to the method of Carlquist (1982) in order to obtain good sections of woody tissues along with Cariquist —Anatomy of Petiveria and Rivina 385 meristematic cells (Fig. 1, 2). The other specimens were sectioned on a sliding micro- tome. Sections were stained with a safranin-fast green combination (a tannic acid— ferric chloride pretreatment was used on the sections of liquid-preserved Petiveria to intensify staining). Macerations were prepared by means of Jeffrey’s Fluid (Johansen 1940) and stained with safranin. Terms accord with definitions of the IAWA Committee on Nomenclature (1964) except for “successive cambia”, which is defined as in Carlquist (1988), which is much like Pfeiffer’s (1926) “Folgemeristeme ausserhalb des primären Cambiumrings sukzessiv von innen nach aussen.” Vessel diameter was measured as lumen diam eter (a mean diameter was estimated for vessels oval in transectional outline). Vessels per group is calculated on the basis that a solitary vessel = 1, a pair of vessels in con tact = 2, etc. Data is based on stems unless otherwise stated.

RESULTS

Petiveria alliacea (Fig. 1—6) Stems and roots with a main cylinder of secondary xylem and secondary phloem, outside of which is a thinner cylinder of secondary xylem and secondary phloem produced by a second (successive) cambium (Fig. 1—3). Mean number of vessels per group, 1.84. Grouped vessels most commonly in radial multiples (Fig. 3); a few ves sels are so narrow that they simulate fiber-tracheids in diameter. Mean vessel lumen diameter, 18 ji.m. Mean number of vessels per mm2, 157. Mean length of vessel ele ments, 170 jim. Mean vessel wall thickness, 2.3 jim. Perforation plates simple, bor ders absent. Lateral wall pits of vessels with circular pit cavities about 5 jim in diam eter; pit apertures very narrowly elliptical. Imperforate tracheary elements are all fiber-tracheids with vestigial pit borders; pit cavities about 2 jim in diameter. Starch present in fiber-tracheids. Mean length of fiber-tracheids, 442 jim. Mean wall thick ness of fiber-tracheids, 2.5 jim. Axial parenchyma vasicentric, forming an incomplete sheath no more than one cell thick around vessels or vessel groups. Axial parenchyma strands composed of two to three cells with lignified cell walls. Multiseriate rays more common than uniseriate rays (Fig. 4). Mean multiseriate ray height, 288 jim. Mean width of rays, 3.2 cells. Mean height of uniseriate rays, 59 jim. Upright and square cells more common than procumbent cells. Ray cell walls about 1.2 jim thick, lignified. Starch present in ray cells. Growth rings present (Fig. 5). Secondary phloem with large longitudinally-oriented styloids (Fig. 3, 5, white areas in the dark gray crushed secondary phloem). Packets of elongate crystals like coarse raphides also present in secondary phloem. The second cambium originates in the cortex, midway between periderm and sec ondary phloem derived from the first cambium. Secondary phloem and secondary xylem are derived from the second cambium, but exterior to the secondary phloem there are radial lines of cells that are apparently derived from the second cambium (Fig. 2); these cells contain starch (more in cells to the outside, which suggests the outermost cells are not meristematic). 386 IAWA Journal, Vol. 19 (4), 1998

, .7

1jj2-’ 2

Fig. 1—4. Petiveria alliacea. — 1 & 2: Stem transections showing second cambium and its products, from Heidelberg Botanic Garden specimen. 1: Cambial zone with little mature sec ondary xylem and phloem. 2: Secondary xylem and secondary phloem (left) and probable ray (right) produced from second cambium; radial files of cells extend outward from the second ary phloem to the point indicated by the arrows. — 3 & 4: Sections from dried specimen, MADw 35452. 3: Transection; secondary xylem and phloem from first and second cambia; the two bands of crushed secondary phloem are darker gray and contain white spaces that are styloids in transection. 4: Tangential section, rays narrow multiseriate and uniseriate. — Fig. 1, 3, 4, magnification scale above Fig. 1; Fig. 2, scale above Fig. 2 (divisions in scales = 10 pm). Carlquist Anatomy of Petiveria and Riina 387

Fig. 5 & 6. Petiveria alliacea, MADw 35452. — 5: Transection of secondary phloem (above) and secondary xylem (below) from second cambium: latewood in bottom third of photograph.

— 6: Radial section of secondary phloem to show styloids in longitudinal view. — Fig. 5 & 6, magnification scale above Fig. 2.

Rivina humilis (Fig. 7—11). A second cambium (and therefore a second cylinder of secondary phloem and sec ondary xylem) not present in my material, but reported by Solereder (1908) and Pfeiffer (1925). Mean number of vessels per mm2, 190. Mean vessel element length, 176 pm. Mean vessel wall thickness, 2.2 pm. Vessels with simple, nonbordered perforation plates. Lateral wall pitting composed of alternate circular bordered pits, pit cavities about 5 pm in diameter, pit apertures narrowly elliptical. Imperforate tracheary ele ments all fiber-tracheids with vestigially bordered pits. Mean fiber-tracheid length, 412 pm. Mean fiber-tracheid wall thickness, 2.5 pm (Fig. 9, 11). Starch present in fiber-tracheids (Fig. 10). Axial parenchyma vasicentric, forming a sheath no more than one cell thick around vessels or vessel groups (Fig. 9). Axial parenchyma in strands of one or two cells, with thin lignifled walls. Multiseriate rays more common than uniseriate rays (Fig. 8). Mean height of inultiseriate rays, 377 pm. Mean width of multiseriate rays, 2.8 cells. Mean height of uniseriate rays, 111 pm. Upright, square, and procumbent cells equally frequent in multiseriate rays. Ray cell walls thin, lignified (Fig. 10, 11). Starch present in ray cells (Fig. 10). Wood nonstoried. “Small crystals F,

388 IAWA Journal, Vol. 19 (4), 1998

H

Fig. 7—il. Rivina humilis wood sections. — 7: Transection, showing narrow diameter of ves

sels. — 8. Tangential section; narrow multiseriate and uniseriate rays are present. —9: Transec

tion to show thin-walled nature of fibers, vasicentric axial parenchyma. — 10: Tangential sec

tion to show starch in ray cells (right) and in fiber-tracheids (left). — 11: Transection of pith and adjacent secondary xylem; a few styloids in transection in pith; two styloids in transection in

the ray are indicated by arrows. — Fig. 7, 8, magnification scale above Fig. 1; Fig. 9, 11, scale above Fig. 2; Fig. 10, scale above Fig. 10 (divisions = 10 j.im). Carlquist Anatomy of Petiveria and Rivina 389

resembling styloids in certain cells of the bast” (Solereder 1908: 666), but not ob served in my material because secondary phloem was not preserved. Styloids present in parenchyma of pith margins, as well as in adjacent areas of secondary rays (Fig. 11, arrows).

CONCLUSIONS

As the above data show, there are numerous similarities between Petiveria and Rivina with respect to stem and wood anatomy. Both genera have a second cylinder of sec ondary xylem and secondary phloem, formed from a cambium, From both the first and the second cambium, secondary phloem rich in large styloids plus packets of elongate crystals like coarse raphides are produced. In both the first and the second increments of secondary xylem, vessels have nonbordered simple perforation plates and alternate bordered pits 5 jim in diameter; quantitative vessel features (vessel group ing, vessel lumen diameter, vessel element length, vessel density); starch-bearing fiber tracheids with vestigial bordered pits are present; axial parenchyma is scanty vasicentric; rays are mostly (narrow) multiseriate, with upright cells predominant and with lignified cell walls. The above list of similarities is rather amazing. One could account for the similari ties by hypothesizing that conspicuous differences between the two genera (e.g., fruit type) are autapomorphic, but that the genera share numerous synapomorphies and thus are closely related. This latter conclusion would be in accord with the adjacent placement of the two genera in Heimerl’s (1934) monograph. Some of the similar character states cited above (e.g., styloids plus coarse raphides in the secondary phloem) are quite unusual. The alternative explanations are that a large number of similarities in wood have been evolved separately in the genera (presumably related to similari ties in habit) and are thus homoplasies; or that similarities represent symplesiomorphies, retained from ancestral Phytolaccaceae s.l. One likely symplesiomorphy is the ab sence of borders on perforation plates in Petiveria and Rivina; this character occurs in other Phytolaccaceae such as Agdestis (Carlquist, in press) although it is uncommon in dicotyledons at large. An assessment of wood diversity in relation to systematics and phylogeny will be attempted in a concluding paper of this series on woods and stems of Caryophyllales. One can compute, from vessel data, a Mesomorphy Ratio for the two genera (ves sel lumen diameter times vessel element length divided by vessel density). The Ratio value for the two genera is the same (19), a figure that would correspond to the Meso morphy Ratio for desert shrubs of southern California (20.9) in the sample of Carlquist and Hoekman (1985). The desert shrub value was computed using outside diameter of vessels rather than vessel lumen diameter but the numerical difference for the Mesomorphy Ratio using the two different methods is not great. The low Mesomor phy Ratio for Petiveria and Rivina is in accord with their habitats in dry tropical and subtropical areas. A shortlived perennial herb draws moisture only from surface soil layers likely to dry more rapidly than deeper soil, especially in open areas where weedier plants like Petiveria and Rivina tend to grow. 390 IAWA Journal, Vol. 19 (4), 1998

Petiveria shows clearly the origin of the second cambium in the cortex. This is in accord with origin of the second cambium as reported by numerous authors who have studied successive cambia (see Pfeiffer 1926). However, the way in which a third cambium originates in dicotyledons with successive cambia and which tissues are produced from the third cambium and how are much more controversial (Gibson 1994; Cariquist, in press). Phytolaccaceae s.l. and other Caryophyllales with succes sive cambia are of key importance in attempting to clarify the details of this phenom enon in dicotyledons at large.

ACKNOWLEDGEMENTS

Dr. Heinz-Dietmar Behnke kindly provided the material of Petiveria alliacea from the Heidelberg Botanic Garden. Dried material of this species was provided by Dr. Regis Miller of the U.S. Forest Products Laboratory, Madison, Wisconsin.

REFERENCES

Behnke, H.-D. 1997. Sarcobataceae — a new family of Caryophyllales. Taxon 46: 495—507. Brown, G.K. & G.S. Varadarajan. 1985. Studies in Caryophyllales. I. Classification of Phytolaccaceae s.l. Syst. Bot. 10: 49—63. Carlquist, S. 1982. The use of ethylene diamine in softening hard plant structures for paraffin sectioning. Stain Techn. 57: 311—317. Carlquist, S. 1988. Comparative wood anatomy. Springer Verlag, Heidelberg & Berlin. Carlquist, S. 1995. Wood anatomy of Caryophyllaceae: ecological, habitat, systematic, and phylogenetic implications. Aliso 14: 1—17. Carlquist, S. 1997. Wood anatomy of Portulacaceae and Hectorellaceae: ecological, habital, and systematic implications. Aliso 16: 137—153. Carlquist, S. In press. Wood, bark, and stem anatomy of Basellaceae with relation to systemat ics and cambial variants. Flora. Carlquist, S. In press. Wood anatomy of Agdestis (Caryophyllales): systematic position and nature of the successive cambia. Aliso. Carlquist, S. & C.J. Boggs. 1996. Wood anatomy of Plumbaginaceae. Bull. Torrey Bot. Club 123: 135—147. Carlquist, S. & D.A. Hoekman. 1985. Ecological wood anatomy of the woody southern Cali fornia flora. IAWA Bull. n.s. 6: 319—347. Downie, S.R., D.S. Katz-Downie & K.-J. Cho. 1997. Relationships in the Caryophyllales as suggested by phylogenetic analyses of partial chloroplast DNA 0RF2280. Amer. J. Bot. 84: 253—273. Downie, S. R. & J. D. Palmer. 1994. A chioroplast DNA phylogeny of the Caryophyllales based on structural and inverted repeat restriction site variation. Syst. Bot. 19: 236—252. Gibson, A.C. 1994. Vascular tissues. In: H.-D. Behnke & T.J. Mabry (eds.), Caryophyllales. Evolution and systematics: 45—74, Springer Verlag, Heidelberg & Berlin. Goldblatt, P., J.W. Nowicke, T.J. Mabry & H.-D. Behnke. 1976. Gyrostemonaceae: status and affinity. Bot. Not. 129: 201—206. Heimerl, A. 1934. Phytolaccaceae. In: A. Engler & K. Prantl (eds.), Die naturlichen Pflanzen familien. ed. 2, 16c: 135—164. Wilhelm Engelmann, Leipzig. Hershkovitz, M.A. 1991. More Centrospermae. II. The phytolaccaceous alliance and the chen ams. Amer. J. Bot. 78 (6: Suppl.): 191 (Abstract). Cariquist Anatomy of Petiveria and Rivina 391

IAWA Committee on Nomenclature. 1964. Multilingual glossary of terms used in describing wood anatomy. Konkordia, Winterthur. Johansen, D.A. 1940. Plant microtechnique. McGraw Hill, New York. Manhart, J. R. & J. H. Rettig. 1994. Gene sequence data. In: H.-D. Behnke & T. J. Mabry (eds.), Caryophyllales. Evolution and systematics: 235—236. Springer Verlag, Heidelberg & Ber lin. Pfeiffer, H. 1926. Das abnorme Dickenwachstum. In: K. Linsbauer (ed.), Handbuch der Pflan zenanatomie 11(2), 10: 1—272. Gebruder Borntraeger, Berlin. Rauh, W. & K. Dittmar. 1970. Weitere Untersuchungen an Didiereaceen. 3 Teil. Vergleichend anatomische Untersuchungen an den Sprossachsen und den Dornen der Didiereaceen. Sitz. Heidelberger Akad. Wiss. 1969/70 (4): 1—88. Rettig, J.H., H.D. Wilson & J.R. Manhart. 1992. Phylogeny of the Caryophyllales — gene sequence data. Taxon 41: 201—209. Rodman, J.E., KG. Karol, R.A. Price, E. Conti & K.J. Sytsma. 1994. Nucleotide sequences ofr rbcL confirm the capparalean affinity of the Australian endemic Gyrostemonaceae. Austral. Syst. Bot. 7: 57—69. Rodman, J.E., M. K. Oliver, R.R. Nakamura, J. U. McClammer & A.H. Bledsoe. 1984. A taxo nomic analysis and revised classification of Centrospermae. Syst. Bot. 9: 297—323. Solereder, H. 1908. Systematic anatomy of the dicotyledons (transl. by L.A. Boodle & F.E. Fritsch). Vol. 2. Clarendon Press, Oxford. Takhtajan, A. 1987. Systema Magnoliophytorum. Oficina Editoria ‘Nauka’, St. Petersburg. Thorne, R.F. 1992. Classification and geography of the flowering plants. Bot. Rev. 58: 225— 348. Tobe, H. & P.H. Raven. 1991. The embryology and relationships of Gyrostemonaceae. Aus tral. Syst. Bot. 4: 407— 420.