WOOD ANATOMY AND RELATIONSHIPS OF PENTAPUYLACACEAE: SIGNIFICANCE OF VESSEL FEATURES

SHERWIN CARLQUIST

Reprinted from PHYTOMOKPHOLOGY Vol. 34, Nos. 1-4, March-December 1984 Reprinted from PHYTOMORPHOLOGY, Vol. 34, Nos. 1-4, March-December 1984 Issued March 31, 1985, pp. 84-90

PRINTED IN SNDIA

WOOD ANATOMY AND RELATIONSHIPS OF PENTAPHYLACACEAE: SIGNIFICANCE OF VESSEL FEATURES’

SHERWIN CARLQUIST Claremont Graduate School, Pomona College, and Rancho Santa Ana Botanic Garden, Claremont, California 91711, U.S.A.

Abstract Three wood samples of Pentaphylax, representing a range of habitats and ages, were examined. They agree in qualitative features: vessels angular in transectional outline; perforation plates scalariform and long; vessel elements long; intervascular pitting scalariform; vessel to-tracheid pits alternate and sparse; vessel-ray pitting scalariform; spirals present in vessel element tips (one collection); pits on tracheids fully bordered; axial parenchyma diffuse to diffuse-in-aggre gates; rays beginning as uniseriates but becoming multiseriate with age (Heterogeneous I of Kribs); dark-staining deposits in rays; and a few crystals in axial parenchyma (one collection). These features are those of a primitive wood, and seem to place the family in Theales along with Clethraceae, Cyrillaceae, and Theaceae; such thealean families as Aquifoliaceae are also close. The ecological correlations of vessel diameter, vessel element length, and number of vessels per sq. mm are discussed. All of these are doubtless valid as indicators of xeromorphy or mesomorphy. The mesomorphy index indicates that the wood of Pentaphylax collections covers a wide range for a single species.

Wood anatomy of Pentaphylax has in the genus, which is the only genus been presented in a condensed form of the family). by Heimsch (1942) and Metcalfe & Chalk Pentaphylacaceae have generally been (1950). The present study was under conceded to be related to Theaceae by taken because the available materials recent authors, but many earlier syste of Pentaphylax wood showed appre matists had placed it in Celastrales (see ciable diversity for a single species (the van Steenis 1955), where it was grouped treatment of the family offered by van with Aquifoliaceae, Corynocarpaceae, and Steenis 1955) Celastraceae. Aquifoliaceae, probably, The diversity of the wood samples should be placed near Theaceae, as was offers a problem for interpretation in perceptively done by Thorne (1968). ecology and ontogeny. Also the wood Corynocarpaceae and Celastraceae are anatomy of these species could probably not close to each other, and have affinities suggest whether Pentaphylax is composed elsewhere. Thus, Pentaphylacaceae can of a single species, as claimed by van be regarded as thealean no matter which Steenis (five species have been named view is considered. Study of wood 1. Received for publication: September 8, 1983. The writer is grateful to the Keeper of the Herbarium of the University of California, Berkeley, for the wood sample from Taam 2251, and to the Forestry Research Institute of Malaysia, Kepong, for the specimen of unknown provenance. 84 CARLQUIST — WOOD ANATOMY OF PENTAPHYLAX 85 anatomy is, however, of potential impor structures (long vessel elements break tance in confirming this. easily in macerations) were present.

Material and Methods Observations

Wood of Pentaphylax euryoides CARLQuIST 4387 — Vessels angular in Gardner & Champion, was collected transection, mostly solitary (Fig. 1). in the Malayan Highlands (Gunung Ulu Mean vessel diameter, 65 m. Mean Kali, also known as Genting Highlands), number of vessels per sq. mm = where it is a dominant element in the 103. Mean vessel wall thickness, 2.5 cloud forest. A wood sample, 8 cm i.m. Vessels with long scalariform per in diameter, representing a reasonably foration plates, mean number of bars mature wood pattern, was taken. Material per perforation plate = 37. Perforations from Hong Kong (Taam 2251) was narrowly bordered in middle of bars, less mature (6 mm in diameter) and was more fully bordered at ends of bars (Figs. obtained from an herbarium specimen. 3,4). Some perforation plates have forked It represents an unusual ecological site bars or a meshwork-like pattern (Fig. 5). for the species, Although van Steenis Vessel-ray pitting scalariform (Fig. 7). (1955) mentions Hong Kong in the dis Vessel-to-vessel pitting scalariform but tribution of Pentaphylax, he does not mostly limited to tips of vessel elements take that locality into account in his because vessels are almost all solitary. elevational distribution (mountain forests Vessel-to-tracheid pitting sparse, alter and subalpine scrub forest, 1200-3000 m), nate. No spirals seen in tips of vessel for the Hong Kong specimen is from elements. Mean vessel element length, near sea level (Repulse Bay Road); his 1189 m. All imperforate tracheary ele distribution map also excludes Hong ments are traëheids with fully bordered Kong from the range of the genus. The pits (Fig. 6). Mean tracheid length, third sample was provided by the Forest 1456 m. Mean tracheid diameter, 30 Research Institute of Malaysia at m. Mean tracheid wall thickness, 9 Kepong. Although the Institute could m. Axial parenchyma diffuse, with not offer information on the provenance some tendency to diffuse-in-aggregates of this wood sample, it seems (by (Fig. 1), parenchyma in strands of 6-9 virtue of angle of rays) to come from cells. Rays predominantly uniseriate; a a fully mature and relatively large few rays 2-3 cells wide (Fig. 2). Mean trunk. height multiseriate rays, 551, m. Mean All the samples were available in dried height uniseriate rays, 645 m. Some form, and were boiled prior to section pits on ray cells bordered (Fig. 7). Ray ing. The specimens, Carlquist 4387 and cells erect to procumbent, the latter Taam 2251, were sectioned on a sliding mostly in the central portions of rays, microtome; the former was treated briefly even if the rays are uniseriate (Hetero with ethylene diamine (Kukachka 1977). geneous Type I of Kribs, 1935, although The material from Kepong proved diffi this type is more typically demonstrated cult because it contained large thin- in the Kepong collection below, where walled vessels mixed with very thick- multiseriate rays are more abundant). walled tracheids. This problem was Dark-staining compounds in ray cells solved by means of the technique sug (Figs. 2, 4) and in small amounts among gested earlier (Cariquist 1982a), with tracheids (Fig. 6). Crystals absent. longer than normal treatment in ethylene Wood nonstoried. diamine. Sections of all coFecti ns were KEP0NG COLLECTION — Features the stained with satranin. Macetiraons were same as in Cariquist 4387, except prepared by means of Jeffrey’s Fluid as follows: Mean vessel diameter, and stained with safranin. Quantitative 121 m. Mean vessel wall thickness, data are based on 25 measurements per 3 m. Mean number of bars per per feature, except where the structure was foration plate = 53. Mean vessel ele infrequent or where only a few intact ment length, 2221 m. Mean tracheid 86 PHYTOMORPHOLOGY [March-December

7

:

Figs. 1-7 1984] CARLQUIST — WOOD ANATOMY OF PENTAPHYLAX 87 length, 3143 m. Mean tracheid dia the occurrence of growth rings in temper meter, 48 m. Mean tracheid wall thick ate species; spirals in vessel element ness, 14 m. Rays both multiseriate tips occur in both families (Giebel & and uniseriate, but the multiseriate rays Dickison 1976). These authors discuss are more common (Fig 8). Mean mul relationship of Clethraceae to Ericaceae. tiseriate ray width = 3-4 cells. Mean Although that relationship is not ruled multiseriate ray height, 1542 m. Mean out by wood anatomy, both Clethraceae uniseriate ray height, 676 m. Erect and Pentaphylacaceae match Theaceae ray cells compose the uniseriate rays more closely (cf. Metcalfe & Chalk 1950), and the uniseriate wings of the multi- and the grouping of Pentaphylacaceae senate rays; the central portions of the with Clethraceae and Cyrillaceae in multiseriate rays are composed of pro Theales (Thorne 1968) seems consonant cumbent cells with a few erect sheathing with information from wood anatomy. cells (Fig. 8). Dark-staining deposits not Wood anatomy of Aquifoliaceae, des observed. cribed in detail by Baas (1973), can also

TAAM 2251 — Some fluctuation be cited in support of Thorne’s concept in vessel diameter and abundance of Theales, for he includes Aquifoliaceae indicative of mild growth ring in that order and the wood of Aqui activity evident (Fig. 9). Mean vessel foliaceae proves, in sum, more thealean diameter, 28 m. Mean vessel wall than similar to wood of other orders. thickness, 12 sam. Mean number of Wood ofPentaphylacaceae isveryprimitive bars per perforation plate== 31, bars compared to that of dicotyledons at narrowly bordered (Fig. 10). Spirals large. Naturally there is a tendency present in tips of vessel elements (Fig. 11). to compare such a wood with equally Scalariform intervascular pitting ob primitive woods in dicotyledons, some served. Mean vessel element length, 829 of which may not be close to Penta ram. Mean tracheid length, 865 m. phylacaceae. Rays which begin as Mean tracheid diameter, 18 m. Mean u niseriate, predominantly or exclusively, tracheid wall thickness ranges from 2-4 but become multiseriate with increase related to growth ring activity. Rays in stem diameter, as in Pentaphylax, almost all uniseriate. Mean uniseriate are known in random dicotyledonous ray height, 387 m. Ray cells mostly families. Some of these are families erect, procumbent cells tending to be with primitive wood features, such as present in central portions of rays. Some Illiciaceae(Carlquist 1982b) and Buxaceae dark-staining deposits present in ray (Cariquist 1982c). Rays with this onto- cells. Crystals seen in a few axial genetic pattern are probably more wide parenchyma cells (Fig. 12). spread than presently reported because many comparative studies do not take Discussion ontogenetic stages into account. INTERPRETATION OF VESSEL ELEMENT RELATIONSHIPS OF PENTAPHYLACACEAE FEATURES The genus Pentaphylax pre The wood of Clethraceae resembles sumably consists of a single species, in that of Pentaphylacaceae,differing only in agreement with van Steenis (1955). The

Pigs. 1-7—- Wood sections (Carlquist 4387; RSA). Pig. 1. Transection; axial parenchyma is diffuse and diffuse-in-aggregates. Fig. 2. Tangential section; most ray are uniseriate. Pigs. 3-5. Perforation plates from radial sections. Fig. 3. Typical perforation plate. Fig. 4. Portion of plate showing presence of borders at ends of perforations. Fig. 5. Meshwork-like perforation plate por tion. Fig. 6. Transection, showing borders on pits of tracheids. Fig. 7. Ray cells from radial section, illustrating scalariform vessel-ray pitting. Figs. 1, 2. Magnification scale above Fig. 1 (finest divisions = 10 gm). Pig. 3. Magnification scale above Fig. 3 (divisions = 10 m). Figs. 4-7. Magnification scale above Fig. 4 (divisions = 10 gm). 88 PFIYTOMORPIIOLOGY March-December

[4 -‘4

iE

Figs. 8-12 — Fig. 8. Tangential section (Kepong, s.n.) showing multiseriate rays. PIgs. 9-12. Sections of Taam 2251 (UC). Fig. 9. Transection; vessels narrow. Pig. 10. Perforation plate from radial section; bars are narrowly bordered. Pig. 11. Spirals in tip of vessel element from radial section. Fig. 12. Portion of transection; dark cell, centre, is axial parenchynia cell containing crystal. rigs. 8, 9. Magnification scale above Fig. 1. Figs. 10-12. Magnification scale above Fig. 4, 19841 cARLQU1ST WOOD ANATOMY OF PENTAPHYLAX 89 three collections studied here represent occurs between vessel diameter and vessel different-sized samples, but all the ana element length. Although Baas (1982) tomical differences cannot be ascribed to and Zimmermann (1983) regarded ontogenetic variations: some must have patterns of vessel element length as an ecological basis, and are presumably “functionless trends imposed by corre heritable. The smallest stem is that of lative restraints “, vessel element length Taam 2251, the intermediate is Carlquist could not possibly parallel so precisely 4387, and the largest is the unnumbered vessel diameter and number of vessels Kepong collection. If we range the per sq. mm (three features governed collections in this series, we find that by different factors), and yet be function in anatomical terms also the series corres less. Baas(1982) and Zimmermann (1983) ponds to: increase in vessel diameter; did not consider the hypothesis (Cariquist decrease in number of vessels per sq. mm l982b), based upon Slatyer’s (1976) ob transection; increase in vessel element servation that spread of air embolisms length; increase in tracheid diameter; tends to be localized or minimized by increase in ratio of length between occurrence of perforation plates, even tracheids and vessel elements. simple perforation plates: evidently, the Differences in lengths of conductive spread of an air embolism tends to be cells correspond to those described by stopped at a constriction rather than Bailey & Tupper (1918), in which lengths in the middle of a vessel element. of tracheary elements in secondary xylem According to this idea, if the vessel ele increase with age; Bailey & Tupper ments are shorter, air embolisms are figure more rapid increase for imper more likely to be localized. Conse forate tracheary elements than for vessel quently, just as short vessels offer a form elements, accounting for the increase of safety (Zimmermann 1982), short in the ratio mentioned in the above vessel elements also offer safety. Vessel listing. Changes in the length of tracheary element length does not run parallel to elements with age, are based on an plant height (Carlquist 1975, Baas 1982), increase in the length of fusiform cambial so it cannot be merely a by product of initials, evidently a rather deep-seated particular lengths of mechanical cells: feature, which shifts slowly during growth the basis must be sought elsewhere in of a stem. Ontogenetic change in vessel most cases (longer mechanical cells may diameter is not based on length of fusi be hypothesized to offer greater mecha form cambial initials, but on degree of nical support). Numerous vessels per cell enlargement. The independence of sq. mm of transection are correlated with diameter from length is clearly shown xeric habitats, and offer a form of safety in growth rings of various dicotyledons. (Carlquist 1975, Zimmermann 1983). Narrow vessels are positively corre If vessel diameter, vessel element lated with xeromorphy in any given length, and vessel number per sq. mm are group of dicotyledons which has radiated all related to the degree of safety woods into various habitats (Carlquist 1966). possess under water stress, the three Narrow vessels, more numerous per dimensions (all easily computed) can sq. mm, offer greater safety (Zimmer be combined, for any given species, into mann 1983). The narrow diameter of an index, termed Mesomorphy (Carl the vessels of wood of branches, as quist 1977). The collection Taam 2251 opposed to main stems, has been noted has a mesomorphy value of 54; the collec in several groups (Carlquist 1969, 1980): tion Carlquist 4387 has a value of 750; this feature is mentioned in relation to and that of the Kepong sample is much the much smaller diameter of vessels in higher, 8218. A similar range in meso the collection of Taam 2251, taken from morphy values was found in Pittospo a branch. Vessel diameter and vessel raceae (Carlquist 1981). Both Penta length (not to be confused with vessel phylax and Pittosporum are evergreen element length) are closely correlated shrubs to small trees, and the scope of (Zimmermann 1982) statistically and not variation in the habitats for the two genera morpho1oically. A similar parallel is similar (although Pittosporum can occur

69 1587-1598 Verlag)

functional interpretations; Am. J. Bot. Sap; the Ascent of Germany: (Berlin, Springer

(Illiciaceae): phylogenetic, ecological, and H M Xylem ZIMMERr?a&wI 1983 Structure

and

CARUIUJsT Wood S 1982b anatomy of Publishers) Illicium

Stain sectioning; Tèchn. 57 311-317 don, U.K.: Martinus W. Nijhoff/Dr Junk

hard in softening pp plant structures for in paraffin Wood Anatomy Baas P. 59-70 (Lon ed.

CARLQUIsT The 1982a S use of tomy ethylenediamine angiosperm of New In trees; Perspectives

ceae;

Allertonia 2 355-392 M H ZIMMERMAI’m 1982 Functional xylem ana

CARuuIsT S Wood anatomy 1981 Pittospora of 6(4) A/iso 57-66

Balanopaceae; Allertonia 2 1-246 19 genetic classification the of flowering plants.

CARLQU1ST 1980 S Anatomy and R of F systematics TH0RNE 1968 of Synopsis a putatively phylo

bot. 64 Gdn. 627-637 Flora Malesiana 1 (5) Ser. 121-124

additional species and concepts; Mo. Ann. C 3 STEENIs 0 G Pentaphylaca VAN 1955 ceae;

CARLQUIsT 1977 S Wood anatomy of (London, Onagraceae: U.K.: Academic Press)

California Press) R 0 SLATYER Plant-Water 1967 Relationships,

U.K.: (Berkeley, Evolution; Press) of University

CARuuIsT Ecological 1975 S of Strategies Xylem Dicotyledons; (Oxford, U.K.: The Clarendon

(Campanulaceae); Biotropica 1 47-72 & C R L METcALFE Cuc 1950 Anatomy of

the

CARLQUIsT Wood S anatomy 1969 of Lobelioideae 80 301-307

wood controlling evolution; (2) 6 25-44 A/iso for anatomical studies; Microscopica Acta

tae: a summary, with on comments B KUKAcWCA factors 1977 F Sectioning woods refractory

CARLQuIsT S Wood anatomy 1966 of Composi Gaz. Bot. 96 547-557

Acad. Sci. 54 Arts 149-204 zation the wood in of rays dicotyledons;

gymnosperms, and anglosperms; Proc. Am. A D KRTBs Salient 1935 lines structure of

speciali

the secondary of xylems cryptogams, vascular 8 83-198

in tracheary A cells. I. between comparison to reference taxonomic groupings;

Lilloa

BAILEY I & W W W TuPPER 1918 Size variation and “Terebinthales” Wettstein of with

Martinus W Nijhoff/Dr Publishers) Junk the secondary of the xylem

“Gruinales”

pp

Anatomy 23-58 Bass (London, P ed. U.K.: HEIMscH C 1942 Comparative anatomy JR of

pectives. In New in Pèrspèctives Wood 92 17-26

logical wood anatomy— and history of pers J. Clethraceae; Elisha Mitchell Soc. lent. sc

P BAAs Systematic, 1982 phylogenetic, and eco K & P W GIEBEL Wood C 1976 DICiISON anatomy

phylogenetic significance; 21 Blumea 193-258 Flora .fèna 172 463-491

hex (Aquifoliaceae) and ecological and its correlations with ecology phylogeny; and

BAAs P The wood 1973 anatomical range in CARLQuIsT S Wood anatomy 1982c Buxaceae: of

Literature Cited

not is evident at present. nisms in angiosperms.

the central in portions elements, of vessel the of herbivore many deterrence mecha

spirals the occur should in not tips rence and crystals may be regarded as one

1966; Baas The 1982). why reason creased herbivore pressure, since occur

increase with xeromorphy (Cariquist might look for correlation a with in

xeromorphy, of spirals vessels since in widely dicotyledons. in so, it is If one

Taarn also could be 2251 indication see an this if tendency proves true more

The spirals vessel element in of gation tips more of groups is advised to

mesomorphy value expected. to be is Pittosporum of (Carlquist 1981). Investi

ecological extremes, the lower thus and the 1980) New and Caledonian species

occupied Taam by to is more subject 2251 species: this holds Balanops in (Cariquist

ecological component habitat also. The less or abundant absent more in xeric

geny, is there an evidently independent or present in mesic species of a genus,

collections phylax related have to be onto can been reported to be more abundant

differences rical the among three Penta ralized such on narrow a basis, crystals

some Pentaphylax). Whereas the nume of collection. Although cannot it be gene

much colder habitats drier in and than Crystals are present only the in Taam

90 PHYTOMORPHOLOGY