IAWA Journal, Vol. 14 (4), 1993: 341-357

WOOD AND BARK ANATOMY OF ARISTOLOCHIACEAE; SYSTEMATIC AND HABITAL CORRELATIONS

by

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

Summary httroducti.on Wood of Aristolochiaceae has vessels with Because concepts of relationships of Aris­ simple petforation plates; lateral wall pitting tolochiaceae and of families now thought to of vessels alternate to scalariform; tracheids, be related to Aristolochiaceae have changed fibre-tracheids or libriform fibres present; ax­ markedly in recent decades, evidence from ial parenchyma diffuse, diffuse-in-aggregates, wood anatomy is of special significance. Be­ scanty vasicentric, and banded apotracheal; cause flowers of Aristolochiaceae are trimer­ rays wide and tall, paedomorphic, multiseri­ ous and (except for Saruma) lack petals, the ate only, little altered during ontogeny (new family was included in Santalales until recent­ rays originate suddenly as wid~ multiseriate ly (e.g., Lawrence 1960). However, Saruma rays); ethereal oil cells present in rays; wood has petals, superior ovaries, and apocarpy, as structure storied. All of these features occur well as ethereal oil cells, a characteristic of the in Lactoridaceae and Piperaceae, and support orders Magnoliales, Laurales, and Piperales the grouping of Aristolochiaceae with these of the superorder Magnolianae (= the order families and the nonwoody family Saurura­ Ranales of earlier systems). Appreciation of ceae. Chloranthaceae may be the family next these facts has led to placement of Aristolo­ closest to this assemblage. Druses character­ chiaceae close to magnolioid rather than san­ istically occur in rays of Aristolochia. Tra­ taloid families (see below). The work of Heg­ cheids in Aristolochia may be correlated with nauer (1960, 1989) reinforced this concept. the lianoid habit, although Holostylis, a cau­ At the time LW. Bailey and his coworkers dex perennial thought close to Aristolochia, were studying anatomy of 'Ranales', the san­ also has tracheids. The fibre-tracheids and taloid concept was prevalent, and they did not libriform fibres of Apama and Thottea may elect to study the Aristolochiaceae. Authors be related to the sympodial shrubby habit of as recently as Lawrence (1960) followed this those two genera. On the basis of one species concept. each of Apama and Thottea, the genera differ All current leading phylogenists place Aris­ with respect to wood anatomy. The paedomor­ tolochiaceae in a concept that can be cited as phic ray structure of all genera of Aristolochia­ Magnolianae, a superorder that includes Mag­ ceae suggests an herbaceous or minimally noliales, Laurales, Piperales, and allied orders woody ancestry rather than ancestors with typ­ (Cronquist 1981; Dahlgren 1980; Takhtajan ically woody monopodial habit. Types of bark 1987; Thorne 1992). The treatments of Takh­ structure observed in the species surveyed are tajan and Cronquist are noteworthy in placing briefly characterised. Storied wood structure Aristolochiaceae close to Piperaceae and Chlo­ and presence of druses and ethereal oil cells ranthaceae. In this, they anticipate the results in rays are newly reported for the family. ofLes et al. (1993) and Qiu et al. (1993). Leins and Erbar (1985) regarded Aristolochiaceae Key words: Aristolochiaceae, Lactoridaceae, as a group close to the origin of monocotyle­ Piperaceae, ethereal oil cells, paleoherbs, dons, a concept not markedly different from storied wood structure, vessel restriction the results expressed by Les et al. (1993) and patterns, wood anatomy. Qiu et al. (1993).

1) Address for correspondence: 4539 Via Huerto, Santa Barbara, CA 93110, U.S.A

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 342 IAWA Journal, Vol. 14 (4), 1993

Wood anatomy of Aristolochiaceae is also tained. Therefore, data on bark have been in­ of significance because of the range of habits corporated into the present study. Sections in the family. Although Aristolochia ranges were stained with safranin and fast green. Ma­ from large lianas to small trees to herbaceous cerations were prepared with Jeffrey's fluid perennials with tuberous roots, Apama and and stained with safranin. Xylem in roots of Thottea are shrubs (although sympodial and the species with tuberous roots was too scarce not markedly woody). Holostylis has some­ to macerate successfully. Scanning electron what woody caudices and less woody upright microscope (SEM) preparations were made stems. Asarum, Hexastylis, and Saruma are by mounting paraffin sections on the alumi­ more nearly herbaceous with minimal accu­ num mounts, dissolving the paraffin, and mulation of secondary xylem. The xylary cor­ sputtercoating the sections. relations with these various habits are instruc­ Locality data for the collections studied are tive and help us understand the relationship as follows: Apama siliquosa Lam., Fosberg between habit and wood anatomy in dicoty­ 57358, POM (14 km north of Ratnapura, Sri ledons as a whole. Habital divergences with­ Lanka); Aristolochia californica Torr. (cult. in Aristolochiaceae correspond to the tribes Rancho Santa Ana Botanic Garden, Clare­ recognised by Schmidt (1935) and Gregory mont, Calif., U.S.A.); A. griffithii Hook. f., (1956): Sarumeae (Asarum, Hexastylis, Saru­ Carlquist 8048 (cult. Royal Botanic Gardens, ma), Bragantieae (Apama, Thottea), and Aris­ Kew, U.K.); A. iquitensis O. C. Schmidt, tolochieae (Aristolochia, Holostylis, Eugly­ C. Davidson 5233, RSA (Mishana, 45 km pha). Habital correlations may explain more from Iquitos up Rio Nanay, Loreto, Peru); about configurations of wood anatomy than A. islandica Pfeiff., F. Barkelew 169, POM systematic groupings. (San Benedicto 1., Mexico); A. orbicularis Metcalfe and Chalk (1950) have provided Duchr., G. Hinton 5258, POM (Pungara­ data on wood anatomy of Aristolochia (species bato, Coyuca, Mexico); A. sipho L'Herit., not given); their data on Apama are based on a Carlquist 8059, RSA (cult. Royal Botanic misidentified specimen, and Rao et al. (1992) Gardens, Kew, U.K.); A. triactinia Hook. f., have performed the useful service of describ­ J. Louis 16779, RSA (Yangambi, Congo); ing the wood anatomy of Apama (Thottea) Tw-42251 (Shaba, West Africa); Asarum siliquosa on the basis of correctly determined hartwegii S. Wats., G. Wallace 1092, RSA material. Descriptions of bark anatomy for (Redwood Mt, Kings Canyon Nat. Park, Tu­ the family have not been hitherto presented. lare Co., California, U. S. A.); Holostylis reniformis Duchr., G. Eiten 10034, RSA (63 km N of Gurupi, Goias, Brazil); Thottea gran­ Materials and Methods diflora Rottb., Jumali, Jan. 18, 1966, RSA Aristolochia wood contains large vessels, (MacRitchie Reservoir, Singapore). which would be subject to collapse during Apama and Thottea are both recognised sectioning on a sliding microtome. The high­ here despite the fact that Hou (1981) has sub­ ly parenchymatous stems of Asarum and the merged Apama into Thottea. The two genera roots of Aristolochia islandica and A. orbicu­ differ by little more, in floral morphology, laris (roots were studied in these two species, than the occurrence of one row of stamens in stems in all other Aristolochiaceae) also could Apama versus two in Thottea. However, the not be sectioned satisfactorily on a sliding mi­ wood of the single species of each genus stud­ crotome. Therefore, a method developed ear­ ied here differs appreciably in imperforate lier (Carlquist 1982) was employed; this pro­ tracheary elements and axial parenchyma so vided quite satisfactory vessel sectioning, and that reference to the two under segregate gen­ also has the merit of expanding, at least in eric names is followed in order to point out part, the phloem and parenchyma cells shrunk­ these differences. The nomenclature for the en during drying. Because of this latter fea­ family used by Schmidt (1935) and Gregory ture, and the capability of the technique to pro­ (1956) has been followed; thus, Isoloma is vide good sections of sclerenchyma embedded not recognised for A. griffithii, and Pararisto­ in soft tissue, good sections of bark were ob- lochia is not used for A. triactinia.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 343

Terminology follows that of the IAWA ous if vessels are wider and vessel density is Committee on Nomenclature (1964). Mea­ greater. The high degree of vessel grouping surements of vessel diameters are based on in Asarum hartwegii is in accord with this prin­ diameter of lumen at widest point. Ray height ciple, because this species has axial parenchy­ could not be measured except for Holostylis ma instead of imperforate tracheary elements, reniformis because most rays extend beyond and occurs in a region where freezing might the limits of a section; accurate ray measure­ disable a proportion of the vessels, so that ments could only be obtained by scraping bark vessel grouping would be of selective value. from stems and measuring rays on un sec­ Mean vessel diameter (Table 1, column 2) tioned stems (e. g., see the illustrations for two in stems of Aristolochia ranges from 51 ~m species of Aristolochia in Fisher and Ewers inA. griffithii (Fig. 16) to 174 ~m inA. triac­ 1992, p. 188). Measurements in Table 1 are tinia (Figs. II, 14). Roots of the herbaceous based on 25 measurements except for wall Aristolochia species (A. islandica, A. orbicu­ thickness, which is based on a typical condi­ laris) have narrower vessel diameter (such a tion for each specimen. species is shown in Fig. 15). Vessel diameters in Apama and Thottea are intermediate in size, Results and Discussion whereas those in Holostylis and Asarum hart­ wegii are narrow. Noteworthy in this connec­ Growth rings tion is that the stems of Aristolochia qualify Growth rings are absent in Apama (Fig. 1; as lianoid, whereas those of Holostylis and see also figures in Rao et al. 1992), tropical Asarum qualify as herbs with a minimum of species of Aristolochia (Figs. 11, 25), the secondary growth. One must note that in li­ tuberous roots of herbaceous species of Aris­ anas, one frequently finds an extended range tolochia (Fig. 15), and Thottea. Growth rings of diameters from very narrow to very wide are present in stems of the three temperate spe­ (Carlquist 1985). The collection of A. triac­ cies of Aristolochia studied here, as shown tinia with flattened stems (Fig. 14) has di­ for A. sipho (Figs. 6, 8) andA. griffithii (Fig. morphism in vessel diameter: vessels are much 16). These, plus A. californica, have growth narrower (75 ~m) in the wood at the edge of rings in which the earlywood is distinguished the stem (two vascular areas at top of Fig. 14), by wider vessels and a band of apotracheal much wider (142 ~m) in wood that parallels parenchyma (Fig. 8). In Asarum hartwegii, the flat surfaces of the stem (two lower vascu­ wider vessels occur in earlywood. In the lar areas, Fig. 14). None of the lianoid species lower stem of Holostylis reniformis (Fig. 19), have large numbers of very narrow vessels, axial parenchyma may be more abundant in a condition seen in lianas of some families latewood, and a band of axial parenchyma other than Aristolochiaceae (Carlquist 1985). may be present prior to earlywood initiation. Areas scanned for vessel density measure­ Vessel elements ment (Table 1, column 3) included ray areas Quantitative vessel features are shown in as well as the fascicular areas. Because rays Table 1, columns 1-5. The number of ves­ are notably wide in most Aristolochiaceae, sels per group (Table 1, column 1) is highest vessel density figures are much lower than in Asarum hartwegii, 6.0, and intermediate in for comparable woods with narrow rays. Rays Apama siliquosa (2.1: Figs. I, 3) and Thottea are exceptionally wide in the roots of the two grandiflora (1.8). In Aristolochia and Holo­ herbaceous species; in one of these, A. islandi­ stylis, the figure ranges from 1.1 to 1.5. The ca, vessel density was 189 per mm2 when rays fact that these two genera have a lower degree were excluded. Vessel density is roughly in­ of vessel grouping is considered to represent versely proportional to vessel diameter (Table another example of the principle that in woods 1, columns 2 & 3). Thus, the species with the with tracheids as the imperforate tracheary ele­ widest vessels, A. triactinia, has the fewest ment type, vessel grouping is minimal and is vessels per mm2. A relatively high number of essentially restricted to contacts produced by vessels per mm2 (201) in A. sipho may relate random vessel distributions (Carlquist 1984). to the large numbers of narrow vessels in Such random contacts would be more numer- latewood (Fig. 6).

Downloaded from Brill.com10/03/2021 08:03:29PM via free access w .j:>. ,.j:>. Table 1. Wood characteristics of Aristolochiaceae.

Species Collection 1 2 3 4 5 6 7 8 9 VG VD VM VL VT TL TT AP RH

Apama siliqUDsaLam. Fosberg 57358 POM 2.05 49 56 306 2.4 1115 7.0 DA Us

Aristolochia californica Torr. cult. RSABG 1.52 88 68 209 2.5 418 4.2 D,DA,DG,E usP

A. griffithii Hook. f. Carlquist 8041 RSA 1.17 51 92 284 4.1 469 2.2 D,DA,DG,E usP

A. iquitensis O.c. Schmidt Davidson 5233 RSA 1.10 85 36 284 2.2 451 2.6 D,SV Us

A. islandica Pfeiff. (root) Barkelew 169 POM 1.52 41 189 148 2.4 ? 2.2 D,DA,DG,SV Us

A. orbicularis Duchr. (root) Hinton 5288 POM 1.08 26 ? 146 4.0 ? 5.1 D Us

A. sipho L'Herit. Carlquist 8059 RSA 1.40 67 201 215 2.0 443 3.3 D,DA,DG,E usP

A. triactinia Hook. f. Louis 16779 RSA 1.24 142 17 406 3.0 876 5.1 D,SV usP

A. triactinia Hook. f. Tw42251 1.20 174 14 299 4.6 989 4.4 D, DA,SV usP Asarum hartwegii S. Wats. Wallace 1092 RSA 6.00 38 504 202 PE U ...... Holostylis reniformis Duchr. (base) Eiten 10034 RSA 1.29 35 68 203 2.0 301 4.0 D,DA,PA USP > Downloaded fromBrill.com10/03/2021 08:03:29PM ~ Thollea grandiflora Rottb. Jumali 1-18-1960 RSA 1.80 52 108 381 1.5 842 2.4 SV U ..... 0

Key 10 columns: 1 (VG), mean number of vessels per group; 2 (VD), mean lumen diameter of vessels at widest point, 11m; 3 (VM), mean number of vessels per mm2 (figure 3 for A. islandica excludes ray areas); 4 (VL), mean vessel element length, 11m;5 (VT), vessel wall thickness, 11m; 6 (TL), mean imperforate tracheary element length, 11m; 7 (TT), ~ <: wall thickness of imperforate tracheary elements, 11m; 8 (AP), type of axial parenchyma (D = diffuse, DA = diffuse-in-aggregates, DG = diagonal aggregations, E = earlywood; ~ PA = extensive patches, PE = pervasive, SV = scanty vasicentric); 9 (RH), ray histology (U = upright, S = square, P = procumbent; upper case indicates predominant cell type -.j:>. or types). ~

via freeaccess \0 -\0 W Carlquist - Wood and bark anatomy of Aristolochiaceae 345

Figs. 1-5. Wood sections of Apama and Thottea. - 1-4: Apama siliquosa. 1: Transection; note wide rays and absence of vessel contacts with rays. 2: Tangential section; rays are wide, tall. 3: Portion of transection, to show diffuse-in-aggregates axial parenchyma. 4: Portion of radial section; pits on fibre-tracheids are bordered. - 5: Thottea grandiflora, portion of imper­ forate tracheary element; only the minutest borders are present on pits. - Fig. 1,2: scale above Fig. 1 (divisions = 10 11m); Fig. 3: scale above Fig. 3 (divisions = 10 11m); Figs. 4, 5: scale above Fig. 4 (divisions = 10 lim).

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 346 IAWA Journal, Vol. 14 (4),1993

Figs. 6-9. Wood sections of Aristolochia sipho. - 6: Transection; the origin of a ray is shown near bottom, centre. 7: Tangential section; note storied condition and druses in rays. 8: Portion of transection. The beginning of the growth ring, near bottom, contains vessels of medium size and axial parenchyma. 9: Portion of radial section, to show bordered pits on tracheids. - Figs. 6,7: scale above Fig. I; Fig. 8: scale above Fig. 3; Fig. 9: scale above Fig. 4.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 347

Figs. 10, 11. Wood sections of Aristolochia. -10: A. sipho, portion of tangential section; storied axial parenchyma at left; arrows in ray indicate enlarged starch-containing cells. 11: A. triactinia (Louis 16779), transection, to show large solitary vessels and axial parenchyma celis, basically diffuse but grouped in various ways. - Figs. 12, 13: A. sipho, SEM photos of druses from tangential section. 12: Druse including small number of lesser crystals. 13: Druse containing large number of lesser crystals. - Figs. 10, 11: scale above Fig. 3. Figs. 12, 13: scales at upper left in each represents 1 /lm.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 348 IAWA Journal, Vol. 14 (4),1993

Figs. 14-18. Wood sections of Aristolochia. - 14: A. triactinia (Louis 16779), transection of stem. 15: A. spec. (Aguirre 189, RSA), transection of root. 16: A. griffithii, transection of stem, showing growth rings. 17: A. californica tangential section of stem; upright sheathing ray cells have lignified walls. 18: A. triactinia (Tw 42251), tangential section; dark walls denote ethe­ real oil cells in the ray. - Figs. 14-18: scale above Fig. 1.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 349

Figs. 19-22. Wood sections of Holostylis and Asarum. - 19-21: Holostylis reniformis. 19: Transection, showing patches of axial parenchyma within a fascicular area of secondary xylem. 20: Tangential section; note shortness of vessel elements and absence of uniseriate or narrow multi seriate rays. 21: Portion of tangential section to show bordered pits on tracheids (left) and vessel elements (right). - 22: Asarum hartwegii, vessels from transection to show scalariform pitting (left) and pits that extend laterally around the vessel. - Fig. 19: scale above Fig. 1; Fig. 20: scale above Fig. 3; Figs. 21, 22: scale above Fig. 4.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 350 IAWA Journal, Vol. 14 (4),1993

Figs. 23-27. Transections of bark of Aristolochiaceae. - 23: Apama siliquosa, ring of cortical fibres intact. 24: A. griffithii, cortical fibre ring broken (two segments visible), with parenchy­ ma cells filling in the breaks. 25: A. iquitensis, sdereids fill in breaks in cortical fibre ring (fibre ring portions indicated by arrows); diffuse fibres in secondary phloem. 26: A. triactinia (Louis 16779), druses in phelloderm. - 27: Holostylis reniformis (upper stem), ring of cortical fibres almost intact. - Figs. 23-27: scale above Fig. 1.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 351

Figures for mean vessel element length Imperforate tracheary elements (Table 1, column 4) are remarkably uniform. In Aristolochia and Holostylis, the imper­ Excepting the rather short vessel elements in forate tracheary elements are tracheids in the roots of the two herbaceous species (A. is­ terminology of the lAWA Committee on No­ landica, A. orbicularis) and the caudex of menclature (1964). The relatively large (5-6 Holostylis reniformis, the range in average 11m in diameter) densely placed bordered pits length of vessel elements in the species stud­ on Aristolochia tracheids are shown in Fig. 9. ied is from 215 11m to 406 11m. The shortest In Holostylis reniformis, tracheid pits are sim­ vessel elements (mean = 215 11m) were mea­ ilar in size, but are somewhat less densely sured in the species with clearly storied wood placed than those on Aristolochia tracheids (Figs. 7, 10). (Figs. 20, 21). Apama siliquosa has nonsep­ Vessel wall thickness (Table 1, column 6) tate fibre-tracheids in which the pit apertures is not uniform in the family. Thick walled ves­ are about 5 IJ.ffi in diameter; the pits are moder­ sels characterise A. griffithii (Fig. 16) and A. ately sparse (Fig. 4). Rao et aI. (1992) claim triactinia (Fig. 11), but in A. sipho (Figs. 6, that imperforate tracheary elements in this spe­ 8) and A. iquitensis, vessel walls are about cies are "non septate with simple pits." In my half as thick. material (which otherwise agrees with the de­ Pits on vessel elements are about 5 11m scription of Rao et aI. 1992 for this species), (vertical axis) in Apama, Holostylis, Thottea, borders are consistently present in pits of the and in the roots of the herbaceous species. In imperforate tracheary elements. In Thottea the stems of the lianoid species of Aristolo­ grandiflora, pit cavities are 2 11m in diameter chia, pit cavity diameter (vertical axis) is 7-8 or less, and bear vestigial borders, or none at 11m, except for A. iquitensis, in which the pit all on some pits. The fibre-tracheids (which cavity is 5 IJ.ffi in diameter. may be considered transitional to libriform Pitting is basically alternate in vessels of fibres) in this species are septate. the family (e. g., Fig. 21), although in Asa­ Asarum hartwegii lacks imperforate trache­ rum hartwegii (Fig. 22) the pits are so elongate ary elements: axial parenchyma substitutes laterally that a scalariform pattern is present, for imperforate tracheary elements as the axial or else a pseudoscalariform pattern (alternate background tissue in which vessels are pits elongated nearly the width of a wall face) embedded. This is apparently also true in Sa­ occurs. In many vessels of Asarum hartwegii, ruma (W.C. Dickison, personal communi­ the pit cavities extend around the vessels rath­ cation). er than being confined to a single wall face; Imperforate tracheary elements average thus a kind of annular condition is present. 842 11m or longer in Apama siliquosa, Aris­ Vessel to vessel or vessel to tracheid pitting tolochia triactinia, and Thottea grandiflora is alternate in Aristolochia, but vessel to axial (Table 1, column 6). The great length oftra­ parenchyma pitting is scalariform or transitio­ cheids in A. triactinia may bear no relation­ nal (Fig. 7) and border width is reduced com­ ship to habit, because other species of Aris­ pared to borders in intervascular pitting. In tolochia that are Hanas have much shorter roots of A. islandica, grooves often intercon­ tracheids. nect pits of two or three pit apertures. In the Wall thickness of imperforate tracheary ele­ remaining species studied, intervascular pits ments (Table 1, column 7) is relatively great are circular to elliptical. In stems of A. califor­ in Apama si/iquosa (7.0 11m: Fig. 3). Thottea nica and A. griffithii, elliptically shaped areas grandiflora has imperforate tracheary elements of thin walls surround many pits. No helical with much thinner walls (2.4 11m). Relatively thickenings were observed in vessels of spe­ thick walled tracheids occur in Aristolochia tri­ cies in this study, although Solereder (1908) actinia (Fig. 14) and in the roots of A. orbicu­ reported helical thickenings in A. tomentosa. laris and in a species with a similar habit Only simple perforation plates were ob­ (Fig. 15). In all other Aristolochia species, served in the species of Aristolochiaceae tracheid walls average 4.2 11m or less in thick­ studied. ness.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 352 IAWA Journal, Vol. 14 (4),1993

Axial parenchyma cannot accurately be measured in sectioned Axial parenchyma occurrence is summa­ material except in the caudex (lower stem) of rised in Table 1, column 8. Metcalfe and Chalk Holostylis reniformis; the mean ray height in (1950) report vasicentric axial parenchyma in this collection is 555 lim. In the 'shrubby' Apama and Aristolochia. In the case of Apa­ Aristolochiaceae, the mean width of rays at ma, their material was misidentified, as has widest point, in cells, is great: 10.9 in Apama been shown by Rao et al. (1992). Apama siUquosa (Figs. 1, 2), 9.8 in Holostylis reni­ siliquosa has axial parenchyma distributed jormis, 6.3 in Thottea grandiflora. Ray width primarily as diffuse-in-aggregates (Fig. 3). is greater, and varies only a little within the Thottea grandiflora, by contrast, has scanty lianoid species of Aristolochia: 17.7 cells in vasicentric axial parenchyma. Axial parenchy­ A. californica (Fig. 17), 14.5 in A. iquiten­ ma is primarily scanty vasicentric in Aristo­ sis, 13.3 in A. sipho (Figs. 6,7), and 15.1 lochia iquitensis, but a few diffuse cells are in A. triactinia, Tw-42251 (Fig. 18). The also present. In the species of Aristolochia tuberous roots of herbaceous species of Aris­ other than A. iquitensis, axial parenchyma is tolochia have narrow plates of vessels and tra­ present as diffuse-in aggregates in addition to cheids separated from each other by extreme­ diffuse cells; aggregations of diffuse cells can ly wide rays (Fig. 15). The report of narrow also run in diagonal directions in A. triactinia multi seriate and uniseriate rays in Apama (Fig. 11), A. californica, and A. iquitensis, (Metcalfe and Chalk 1950) is based on mis­ as well as in Holostylis reniformis (Fig. 19). identified material, as shown by Rao et al. Irregular patches of axial parenchyma occur (1992). in H. reniformis (Fig. 19). Banded axial pa­ Ray histology is summarised in Table 1, renchyma has been cited above in earlywood column 9. Ray cells are predominantly up­ of species of Aristolochia with growth rings: right in Apama siliquosa (Fig. 2), Aristolo­ A. californica, A. griffithii (Fig. 16), and A. chia iquitensis, A. islandica, A. orbicularis, sipho (Fig. 8). The report of Metcalfe and Asarum hartwegii, and Thottea grandiflora. Chalk: (1950) that axial parenchyma is para­ Predominantly procumbent ray cells, with tracheal in Aristolochia may be a byproduct upright cells as sheath cells on ray margins, of the difficulty of describing a parenchyma were observed in Aristolochia californica distribution that is basically diffuse, but in (Fig. 17), A. sipho (Fig. 7), and A. triactinia which some of the diffusely distributed cells (Fig. 18). touch vessels, as one would expect if axial The wide, tall rays of Aristolochiaceae, parenchyma is distributed randomly. My pref­ their infrequent subdivision, and the lack of erence in a species with diffuse cells is to use uniseriate rays in fascicular areas are, to me, the term paratracheal for axial parenchyma indicative of an herbaceous ancestry for the only if axial parenchyma cells in contact with family because these features are typically vessels are more numerous than one would found in primary rays of herbs. The predom­ expect on the basis of random distribution of inantly upright cells are also characteristic of axial parenchyma cells. paedomorphic rays. One should note predom­ Axial parenchyma in Asarum can be de­ inance of procumbent cells in rays in three scribed as pervasive, because secondary xy­ species of Aristolochia (Table 1, column 9). lem vessels are embedded in a ground tissue Thus, one paedomorphic feature is lacking in of parenchyma, and imperforate tracheary those species. To be sure, the abundance of elements are absent. procumbent cells increases with increase in Axial parenchyma is commonly in strands stem diameter, and the stems studied of the of two cells in Aristolochiaceae (Figs. 7, 10). three species of Aristolochia were larger than Occasional undivided axial parenchyma cells the roots or stems of the remaining taxa. also can be found in any given species. Ray cells have thin nonlignified walls (also more common in herbs and Hanas than in Rays typically woody dicotyledons) in most Aris­ In all Aristolochiaceae studied, rays are ex­ tolochiaceae (Figs. 6, 7, 10, 14-16, 18). clusively multiseriate. Rays are so tall they Sheath cells (upright cells at ray margins)

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 353 have lignified walls in Aristolochia califor­ to the central portions of these fascicular areas nica (Fig. 17) and occasionally in A. sipho. (Fig. 1), and vessels are absent adjacent to Diffusely distributed thin-walled sclereids oc­ ray areas. This pattern is also clear in the il­ cur in rays of A. iquitensis (Fig. 25, lower lustrations of Rao et al. (1992) for this spe­ left). Ray cells are thin walled but uniformly cies, although they do not mention this condi­ lignified in Apama siliquosa (Figs. 1, 2) and tion. Interestingly, there are narrow fascicular Thottea grandiflora. areas in Apama siliquosa, such as the one at Druses in ray cells were observed in all the top of Figure I, in which vessels are ab­ the lianoid species of Aristolochia studied: A. sent. Vessel restriction can take various forms, californica, A. griffithii, A. sipho (Figs. 6, but the original use of the term (Carlquist 7, 12, 13), and A. triactinia (Fig. 14). The 1983) in Valerianaceae is an instance like that druses, as seen with SEM, may incorporate of Apama. Similar instances occur in Ber­ fewer smaller crystals (Fig. 12) or more of beridaceae, Papaveraceae, and Ranunculaceae them (Fig. 13). (Carlquist & Zona 1988). Ethereal oil cells do not occur in rays of Apama siliquosa; reports of these cells in this Bark anatomy species by Metcalfe and Chalk (1950) are the The sampling here of Aristolochiaceae is result of misidentified material (Rao et al. too small to encompass the diversity in bark 1992). However, ethereal oil cells do occur structure that ultimately is likely to be uncov­ in idioblastic fashion in rays of Aristolochia ered for the family. There are, however, some and Holostylis. These cells were observed in unifying patterns. rays of A. californica, A. iquitensis, A. or­ Phloem fibres are absent in stems of Aris­ bicularis, and A. triactinia (Fig. 18). These tolochia californica, A. griffithii, A. sipho, appear as double-walled cells in which the A. triactinia, Asarum hartwegii, and in the thicker inner wall (darker in colour) is often roots of A. orbicularis. Small strands of fibres collapsed and separated from the outer wall. in the protophloem position were observed in In two species studied here, there are enlarged Apama siliquosa (Fig. 23), A. islandica, and ray cells that contain starch: A. griffithii and the lower stem of Holostylis reniformis. In A. sipho (Fig. 10, arrows). These cells might the lower stem of Holostylis reniformis suc­ be idioblasts similar to ethereal oil cells, but cessive cork cambia originating in secondary contents and wall characteristics do not per­ phloem carry these strands outward in the mit them to be designated as ethereal oil cells. phellem the cork cambia produce. Diffuse fi­ bres occur in the secondary phloem of Aris­ Storied structure tolochia iquitensis (Fig. 25) and upper stems Storied wood structure has not hitherto of Holostylis reniformis (Fig. 27). been reported in Aristolochiaceae. However, Druses are present in phloem rays of Aris­ the tracheids and axial parenchyma of Aristo­ to lochia californica, A. griffithii, and A. si­ lochia sipho are clearly storied (Figs. 7, 10). pho. In A. iquitensis, diffusely distributed Aristolochia sipho was the largest of the stems sclereids occur in phloem rays (Fig. 25). studied here, so the appearance of storying is A ring of cortical fibres is characteristic of more likely than in other species (in dicotyle­ stems of all Aristolochiaceae studied, but not dons that have storied wood structure, story­ of the roots. In Apama siliquosa (Fig. 23), this ing becomes more apparent with age). Vague fibre ring is unbroken, but this is a relatively storying was seen in tracheids of some other small stem. Small breaks are visible in the ring Aristolochia species, and the arrangement of of cortical fibres in upper stems of Holo­ ray cells in Apama siliquosa (Fig. 2, especial­ stylis reniformis (Fig. 27). More conspicuous ly near bottom) is suggestive of storying. breaks in the fibre ring, with parenchyma cells filling the breaks, is shown for Aristolochia Vessel restriction patterns griffithii (Fig. 24) and also occurs in A. sipho. The vessels of Apama siliquosa are not ran­ Sclereids formed from the parenchyma cells domly distributed within the fascicular areas enter the breaks in A. californica, A. iquiten­ in which they are located; they are restricted sis (Fig. 25), and A. triactinia (Fig. 26). Soli-

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 354 IAWA Journal, Vol. 14 (4), 1993 tary rhomboidal crystals may be seen in some have termed this assemblage 'paleoherbs', be­ of the sc1ereids that fill in gaps in the broken cause to these workers they appear to have fibre ring in A. triactinia. In the stem figured been primitively herbaceous, as opposed to for A. iquitensis, the strand of fibres shown the woody habit claimed to be basic to many is separated from the sclereids that fill in the dicotyledon groups. I will use the piperalean­ breaks in the fibre ring; such displacement is chloranthalean group for comparisons with unusual. respect to wood anatomy. In a paper on Lac­ Sparse diffusely distributed cortical scle­ toridaceae (Carlquist 1990), I listed ten dis­ reids are present in stems of Apama siliquosa tinctive wood features shared by Lactoridaceae (Fig. 23). The other species studied are devoid and Piperaceae: vessels with simple perfora­ of diffuse cortical sclereids. tion plates; vessels in radial chains or pore Periderm was present in all Aristolochia­ multiples; vessel to axial parenchyma pitting ceae studied, except for the upper stem of scalariform to alternate, vessel-to-vessel pit­ Holostylis reni/ormis (Fig. 27); only a few ting alternate; imperforate tracheary elements pockets of periderm had formed on th~ stem with vestigially bordered pits; imperforate tra­ of Apama siliquosa studied. At least several cheary elements storied, conforming to the layers of phelloderm were present in all Aris­ storied pattern of vessels and axial parenchy­ to lochia stems studied. In the phelloderm of ma; axial parenchyma scanty vasicentric; axial A. iquitensis, oil cells were observed. In phel­ parenchyma cells not subdivided, or, if in loderm of A. cali/ornica and A. triactinia strands of two cells, the septum a thin wall; (Fig. 26), druses occur in phelloderm cells. rays wide and tall, with little ontogenetic sub­ Phelloderm was not observed in the lower division (as opposed to active subdivision of stem of Holostylis reni/ormis, a fact likely rays in most woody dicotyledons); ray cells correlated with the production of successive mostly upright; ray cells taller adjacent to fas­ periderms (likely one periderm per year). cicular areas. Secondary growth in dicotyledons typi­ All ten features just listed can be found in cally results in increased separation between at least some Aristolochiaceae. Fibre-tracheids phloem fibre strands that are already in con­ occur in Apama, but the tracheids of Aristo­ tact with each other. The formation of a ring lochia and Holostylis are more primitive, be­ of cortical fibres, which eventually breaks cause of their more prominently bordered pits with increase in girth (but with sclereids (according to Baileyan theory). Stress should formed within the breaks), seems characteris­ be laid on the presence of storied wood, newly tic of Aristolochiaceae as a whole. Does the reported for Aristolochiaceae, because Aris­ occurrence of a closed sclerenchyma ring in­ tolochiaceae, Lactoridaceae, and Piperaceae dicate an herbaceous ancestry, or is this mere­ are the only families of the superorder Mag­ ly a characteristic of aristolochiaceous bark? nolianae (Thorne 1992; also known as Mag­ Further studies may illuminate this question. noliiflorae or Annonanae) in which storying One sees in transections of stems of herba­ is known except for Berberidales (Papave­ ceous dicotyledons encasement of the vascu­ rales), an order that appears further removed lar cylinder or even of individual bundles by from the piperoid families. fibres; these fibres appear related to lack of Ethereal oil cells (reported as 'mucilage and secondary growth that would break these oil cells' for rays of 'some species of Piper' sheaths open, hence the confinement produced (e.g., P. elongatum Vah!) are known for Pi­ by a fibrous cylinder may cpnnote an herba­ peraceae (Metcalfe & Chalk 1950). Newly ceous mode of stem structure. reported here are oil cells in rays of four spe­ cies of Aristolochia and rays of Holostylis Systematic conclusions (liquid-preserved material of three Aristolo­ Molecular data place Aristolochiaceae close chia species suggests these are oil cells rather to Lactoridaceae in a clade that also includes than mucilage cells, because mucilage rem­ Piperaceae and Saururaceae (Les et al. 1993) nants could not be seen). Qiu et al. 1993). Chloranthaceae and mono­ Vasicentric (or scanty paratracheal) paren­ cotyledons are outgroups of this clade. Some chyma is not the sole parenchyma type for

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 355

Aristolochiaceae, although it does characterise tracheary elements with borders on pits much wood of Aristolochia iquitensis and Thottea reduced. In the tribe Aristolochieae (Aristo­ grandiflora. Note should be taken, as in the lochia, Holostylis), tracheids with prominent­ discussion above, of the tendency for some ly bordered pits are the imperforate tracheary workers to say vasicentric parenchyma is elements. This feature may indicate a close present if some diffuse cells touch vessels relationship between Holostylis and Aristo­ (the term vasicentric or paratracheal is best lochia, particularly since the habit is so differ­ restricted to instances in which diffuse cells ent in these two genera. are absent or appreciably fewer than paren­ Differences in wood anatomy among Lac­ chyma cells in contact with vessels): in any toridaceae, Piperaceae, and Aristolochiaceae random distribution of diffuse axial paren­ are sufficiently minor so that on the basis of chyma, some of these cells are likely to touch that evidence alone, Aristolochiaceae could vessels. Diffuse or diffuse-in-aggregates axi­ be included in Piperales. This is in line with the al parenchyma in Aristolochiaceae frequently cladograms produced by Les et al. (1993) and touch vessels. Qiu et al. (1993). Thus, wood anatomy of Aristolochiaceae Apama siliquosa differs from Thottea is amazingly similar to that of Lactoridaceae grandiflora in several respects (conditions of and Piperaceae, justifying the grouping by Thottea grandiflora in parentheses): imperfo­ Les et al. (1993). Saururaceae are not consid­ rate tracheary elements non septate with thick, ered here because they have little secondary 7 J..lm, walls (septate, wall thickness 2.4 J..lm); xylem. Of the families in the clade next closest pits in fibre-tracheids clearly bordered (bor­ to the piperalean clade just cited in the system ders on pits vestigial or absent), pit cavity of Les et al. (1992), all have relatively primi­ diameter 5 J..lm (diameter 2 J..lm or less); axial tive wood except for Nymphaeaceae, which parenchyma diffuse-in-aggregates (scanty va­ lack secondary xylem. Among these families, sicentric). Investigation of other species in Chloranthaceae are of particular interest be­ these genera is desirable to demonstrate cause they have wide, tall paedomorphic rays whether these differences consistently char­ like those of Aristolochiaceae, Lactoridaceae, acterise the respective genera. and Piperaceae. Chloranthaceae also have tra­ Pararistolochia O.c. Schmidt is a genus cheids (Sarcandra) or fibre-tracheids (other segregated from Aristolochia; it includes A. genera) and diffuse parenchyma with transi­ triactinid. This species is highly distinctive tion to scanty vasicentric (Carlquist 1992b). anatomically; it has diffuse sclereids in xylem Chemical data presented by Hegnauer and phloem rays, rhomboidal crystals in cor­ (1989) led him to mention families such as tical sclereids, thick-walled tracheids, relative­ Annonaceae, Monimiaceae, and Menisper­ ly long vessel elements and tracheids, and maceae as possessing compounds close to flattened stems (with dimorphism in vessel those of Aristolochiaceae. These families be­ diameter related to that flattening). Further long to the superorder Magnolianae (Magno­ studies are needed to show how many fea­ liiflorae), and the first two families are closer tures of this species occur in other Pararisto­ to the piperoid families than are Berberidales, lochia species, and therefore whether the seg­ which inCludes Menispermaceae. Berberi­ regate genus should be recognised. Another dales may be worthy of consideration as po­ genus segregated from Aristolochia, [soloma tentially related to the piperalean clade because (represented here by A. griffithii), has wood they have such features as storied wood struc­ within the patterns shown by the species of ture, wide paedomorphic rays, and vessel re­ Aristolochia sensu stricto. striction patterns. Wood anatomy can be cited in relation to Correlations with habit subfamilial taxonomic arrangements, using Aristolochiaceae do not appear to be primi­ the systems of Schmidt (1935) and Gregory tively woody. One can cite two species of (1956). The tribe Asareae (Asarum, Saruma) Aristolochiaceae that are small trees: A. arbo­ lacks imperforate tracheary elements. The tribe rea Linden andA. leuconeura Linden. These Bragantieae (Apama, Thottea) has imperforate presumably monopodial small trees are likely

Downloaded from Brill.com10/03/2021 08:03:29PM via free access 356 IAWA Journal, Vol. 14 (4),1993

only products of adaptive radiation in a large tracheids appear to be stronger mechanically genus: the herbaceous species of Aristolochia (pits sparser) than those of Aristolochia, and that produce shoots of finite duration from a this would be of significance in the upright succulent tuberous root are another product stems of Holostylis. Likewise, absence ofim­ of such radiation. perforate tracheary elements in Asarum would Apama and Thottea are termed 'shrubs' correlate with lack of mechanical strength in for lack of a better term by most authors. the sprawling stems of this herbaceous genus. However, these are not woody shrubs in the The lateral wall pitting of vessels in Asarum sense of leguminous shrubs. Rather, they ap­ (pits laterally widened) also would confer re­ pear to be shrubs with sympodial branching in duced mechanical strength (and greater flexi­ which each branch lasts for several years and bility). is canelike, as in Nandina or Chloranthus. As Does wood of Aristolochiaceae indicate a stressed earlier (Carlquist 1992b), the sym­ woody or an herbaceous habit as primitive podial habit appears basic to Piperales, and for the family? This question is of importance, may be a reason for bringing Chloranthaceae because families included in Piperales and close to Piperales. The sympodial stems of Nymphaeales are now regarded as 'paleo­ monocotyledons may be indicative that their herbs', which would mean that the herba­ origin is close to these families. The shift ceous habit is symplesiomorphic for these from limited woodiness to the shrubby habit clades (although the origin of the dicotyle­ of Apama and Thottea may be correlated with dons could still have been woody). Wide, reduction in borders on imperforate tracheary high rays with an abundance of upright cells elements, since that would produce a mechani­ suggest an herbaceous ancestry (Carlquist cally stronger cell type. This shift (e.g., sep­ 1962). The absence of narrow multiseriate tate fibre-tracheids with minute borders on rays or uniseriate rays in Aristolochiaceae is pits in Hedyosmum) has occurred in Chloran­ curious, as is the origin of new multiseriate thaceae also (Carlquist 1992b). rays not as narrow multiseriate rays but as de Wide rays and presence of tracheids in novo wide multiseriate rays. Both of these Aristolochia correlate with the lianoid or features occur in Lactoridaceae (Carlquist sprawling habit of species in this genus. Fami­ 1990) and Piperaceae (Metcalfe and Chalk lies with lianoid habit more commonly have 1950). Both of these features can be found in true tracheids or vasicentric tracheids than do Cucurbitaceae (e.g., Acanthosicyos, Cocci­ families with arboreal habit (Carlquist 1985). nia), a family that seems primitively herba­ Greater quantities of parenchyma, either as ceous or nearly so, and in which the lianoid rays as in Aristolochiaceae, or as axial paren­ or vining habit is common (Carlquist 1992a). chyma in scandent genera of other families, These may be features of an herbaceous group also characterise lianas, as opposed to trees with wide primary rays, a group in which or shrubs (Carlquist 1985). The notably wide some phylads shifted to a woodier stature. rays of tuberous roots of herbaceous species of Aristolochia, however, doubtless have a References storage function, because the cells in these rays are rich in starch and may store water Carlquist, S. 1962. A theory of paedomor­ also. phosis in dicotyledon woods. Phytomor­ The vessel elements and tracheids of Aris­ phology 12: 30-45. tolochia are relatively short, especially in the Carlquist, S. 1982. The use of ethylenedia­ temperate species. This may reflect the lack mine in softening hard plant structures for of selection pressure for longer (and there­ paraffin sectioning. Stain Techn. 57: 311- fore stronger) imperforate tracheary elements 317. in the self-supporting stems of Apama and Carlquist, S. 1983. Wood anatomy of Caly­ Thottea, as opposed to the lianoid stems of ceraceae and Valerianaceae, with com­ Aristolochia, which are not self-supporting. ments on aberrant perforation plates in Holostylis, a sort of 'caudex perennial', predominantly herbaceous groups of di­ has short vessel elements and tracheids. The cotyledons. Aliso 10: 413-425.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access Carlquist - Wood and bark anatomy of Aristolochiaceae 357

Carlquist, S. 1984. Vessel grouping in dicot­ Hegnauer, R. 1989. Aristolochiaceae. In: R. yledon woods: significance and relation­ Hegnauer, Chemotaxonomie der Pflanzen, ship to imperforate tracheary elements. Volume 8: 75-83. Birkhauser Verlag, Aliso 10: 505-525. Basel. Carlquist, S. 1985. Observations on func­ Hou, Ding. 1981. Flora Malesianae Praecur­ tional wood histology of vines and lianas: sores LXII. On the genus Thottea Rottb. vessel dimorphism, tracheids, vasicentric (Aristolochiaceae). Blumea 27: 301-332. tracheids, narrow vessels, and parenchy­ IAWA Committee on Nomenclature. 1964. ma. Aliso 11: 139-157. Multilingual glossary of tenns used in Carlquist, S. 1990. Wood anatomy and re­ wood anatomy. Verlagsbuchanstalt Kon­ lationships of Lactoridaceae. Amer. J. Bot. kordia, Winterthur, Switzerland. 77: 1489-1505. Lawrence, G.H.M. 1960. Taxonomy of Vas­ Carlquist, S. 1992a. Wood anatomy of select­ cular Plants. MacMillan, New York. ed Cucurbitaceae and its relationship to Leins, P. & C. Erbar. 1985. Ein Beitrag zur habit and systematics. Nord. J. Bot. 12: Blutenentwicklung der Aristolochiaceen, 347-355. einer Vennittlergruppe zu den Monokoty­ Carlquist, S. 1992b. Wood anatomy and stem len. Bot. Iahrb. Syst. 107: 343-368. of Chloranthus; summary of wood anat­ Les, D., D. Crawford, M. Chase, D. Garvin, omy of Chloranthaceae, with comments C. Wimpee & T. Stuessy. 1993. Mole­ on relationships, vessellessness, and the cular evidence for the phylogenetic posi­ origin of monocotyledons. lAWA Bull. tion of Lactoridaceae, an enigmatic island n.s. 13: 3-16. endemic. (Submitted to Systematic Bot­ Carlquist, S. & S. Zona. 1988. Wood anat­ any.) omy of Papaveraceae, with comments on Metcalfe, C.R. & L. Chalk. 1950. Anatomy vessel restriction patterns. IAWA Bull. of the dicotyledons. Clarendon Press, Ox­ n. s. 9: 253-267. ford. Cronquist, A. 1981. An integrated system of Qiu, Y.-L., M.W. Chase, D.H. Les & C.R. classification of flowering plants. Colum­ Parks. In press. Molecular phylogenetics bia Univ. Press, New York. of the Magnoliidae: cladistic analyses of Dahlgren, R.M. T. 1980. A revised system nucleotide sequences of the plastid gene of classification of the angiospenns. Bot. rbeL. Ann. Missouri Bot. Garden 80. J. Linnean Soc. 80: 91-124. Rao, R. V., R. Dayal, B.L. Sharma & L. Donoghue, M.l & lA. Doyle. 1989. Phylo­ Chauhan. 1992. Reinvestigation of the genetic analysis of angiospenns and the wood structure of Thottea siliquosa (Aris­ relationships of Harnamelidae. In: P.R. tolochiaceae).IAWA Bull. n.s. 13: 17- Crane & S. Blackmore (eds.), Evolution, 20. systematics, and fossil history of the Schmidt, O. C. 1935. Aristolochiaceae. In: Harnamelidae. Vol. 1. Introduction and A. Engler & K. Prantl (eds.), Die natiir­ 'lower' Harnarnelidae: 17-45. Clarendon lichen Pflanzenfamilien ed. 2, 16B: 204- Press, Oxford. 242. Verlag Wilhelm Engelmann, Leipzig. Fisher, J.B. & F.W. Ewers. 1992. Xylem path­ Solereder, H. 1908. Systematic anatomy of ways in liana stems with variant secondary the dicotyledons, trans. L. A. Boodle & growth. Bot. l Linn. Soc. 108: 181-202. F. E. Fritsch. Oxford Univ. Press, Ox­ Gregory, M.P. 1956. A phyletic rearrange­ ford. ment of the Aristolochiaceae. Amer. l Takhtajan, A. 1987. Systema Magnoliophy­ Bot. 43: 110-122. torum. Officina Editoria NAUKA, St. Pe­ Hegnauer, R. 1960. Chemotaxonomische Be­ tersburg. trachtungen 11. Phytochemische Hinweise Thome, R. F. 1992. Classification and geo­ flir die Stellung der Aristolochiaceae in graphy of flowering plants. Bot. Rev. 58: System der Dikotyledonen. Die Pharmazie 225-348. 15: 634-642.

Downloaded from Brill.com10/03/2021 08:03:29PM via free access