BULLETIN OF THE TOBEEY BOTANICAL CLUB

VOL. 97, No. 6, pp. 353-361 NOVEMBER-DECEMBER 1970

Wood anatomy of insular species of and the problem of raylessness Sherwin Carlquist1 Claremont Graduate Sehool and Baneho Santa Ana Botanic Garden, Claremont, California 91711 CARLQUIST, SHERWIN (Claremont Graduate School, Claremont, California 91711). Wood anatomy of insular species of Plantago and the problem of insular woodiness. Bull. Torrey Bot. Club 97: 353-361. 1970.—The most markedly woody species of Plantago are insular. Detailed anatomical descriptions, quantitative data, and photomicrographs are presented for woods of P. arborescens (Canary Islands, lowlands), P. maderensis (Ma­ deira, lowlands), P. webbii (Canary Islands, alpme areas), P. fernandeziana (Juan Fer­ nandez Islands, rain forests), and P. princeps (Hawaiian Islands, shady areas in rain forest). Vessel element length and width are greatest in the rain-forest species, least in the alpine species. These expressions seem clearly correlated with degrees of xeromorphy or mesomorphy. Plantago woods are all basically rayless, but rays are eventually formed in larger stems of the species from Macaronesia (the term given collectively to the Cape Verde Islands, Canary Islands, Madeira, Azores, and nearby islands). Rayless woods as exemplified bj' Plantago tend to occur in dicotyledons in which (1) eambial activity is limited or finite; (2) woodiness is phyletieallv in the process of increase, rarlier tha crease; (3) fusiform eambial initials are relatively short; and (4) juvenilism or paedo- morphosis occurs.

The phenomenon of ^insular moodi­ duce during this period. In the Canaries. ness," the tendency" for insular representa­ P. arbort«c< ns can live through the mild tives of a taxonomic group to be more and more humid summers and thus last for woody than their continental counterparts, several years. In fact, in the relatively is well illustrated by Plantago. This world­ maritime climate of Santa Barbara, Cali­ wide genus is herbaceous, yet on various fornia, cultivated of P. arbor•

Plant ago maderensis of Madeira occurs of this species, which may now be extinct, Dear sea level on Madeira, with Euphorbia. was available to me. Aeonium, Lytantkus, and shrubby cruei- The significance of Plantago woods lies fers. It is closely related to P. arborescens in their response by means of increased both in morphology and ecology, and may woodiness to the uniformity of insular cli­ I"' regarded as a vicarious species of the P. mates. These plantagos are, in fact, sensi­ arborescens group. tively adjusted to the ecological conditions riant ago princeps is a distinctive Ha­ in which they grow. They are not, in my waiian endemic. Unlike other Hawaiian opinion, relicts: they are all on volcanic plantagos, it has slender upright stems with islands of modest age. Apparently phyletic relatively long internodes. In habit it re­ increase in vessel-element length to suit calls some species of Dracaena. Xow ex­ mesic conditions can take place; if so, the tremely rare, Plantago princeps once was mechanism is a pertinent subject for in­ abundant in ravines, on cliffs, and near vest igation. waterfalls in the Hawaiian Islands. The Plantago woods are all rayless, at least stems, although upright, lack mechanical earlier in the ontogeny of a . The sig­ strength and the plants tend to lean on sur­ nificance of rayless woods in relation to rounding shrubs, or to grow up through habit, ecology, level of phyletic specializa­ such other plants. The wood sample used tion, direction of phylesis, and nature of in the present study was collected by Mr. wood ontogeny, needs to be understood. Melvern Tessene in the Aina IIo Valley, Data of the present paper are pertinent to Koolau Gap, . This plant would fall this problem. into P. princeps var. la&ifolia according to Wood of Plantago princeps was cited the treatment of Rock (1920). Rock recog­ by the writer (1962) in such a context. A nizes numerous varieties of P. princeps. description of wood of Plantago, based on These might be termed subspecies accord- P. fernandeziana, was offered bv Metcalfe in-.'' to current concepts. I prefer, however. and Chalk (1950). to think of P. princeps as a single polymor­ Materials and methods. Woods of the phic species. In fact, P. princeps is appar­ Maearonesian species of Plantago were col­ ently fully interfertile with Hawaiian lected by the writer in the field. The largest plantagos of an entirely different appear­ plants available were taken for study. Some ance, including bog species such as P. samples were unsuitable for study, how­ pachyphyUa, P. hillebrandii. etc. (Melvern ever, because of markedly distorted wood Tessene. persona] communication). Ecologi- grain—stunted, twisted stems are common eal data and photographs of P. princeps in the Maearonesian species. The writer is and other Hawaiian plantagos are offered indebted to Mr. Melvern Tessene for the by Kock (1920) and Carl.piist (1970a). sample of P. princeps. The samples of P. Plantago fernandeeiana is a species of fernandeziana were provided through the rain forest in the Juan Fernandez Islands. assistance of Dr. Otto Solbrig, Dr. Fred Ms ecological situation is much like that of Meyer, and Dr. Richard H. Eyde. P. princeps. The internodes are shorter Wood sections and macerations were than in P. princeps. but it also tends to prepared according to the usual techniques. form unbranched upright stems topped by The wood samples available were reasonable leaf rosettes. Good photographs of P. fer- in size for a wood study, and samples with tinndi :inIKI and ecological information are too limited a development of secondary ided by Skottsberg (19S2). One can­ x vie in were deliberately omitted. For ex­ not be positive that P. fernamleziana is not ample, liquid-preserved stems of P. pachy- closely related to P. princeps, but it could jdndla could have been sectioned by means AVCII be an insular derivative from main­ of paraffin techniques, but the limited de­ land plantagos independent of other insular velopment of secondary xylem in this spe- Pacific species of Planiago, In the Atlantic. eies would, in my opinion, not have added the distinctive P. rob'uata, an endemic of significant information, at least with re­ St. Helena, is almost certainly not closely spect to the central concerns of the present related to the Maearonesian plantagos. For study. information and illustrations of P. robusta. Qualitative and quantitative data were the reader is referred to Melliss (1875) and obtained from six collections of the five Carlqnist (1965). Unfortunately, no wood species studied here. In rayless or near- L9T0 CARLQUIST: WOOD ANATOMY OF PLANTAGO 355 rayless woods, ray measurements are not jacent to vessels may perhaps be termed axial of significance. The diameter of libriform parenchyma cells by virtue of their pits, which ider apertures than typical for libriform fibers was obtained by measuring the widest fibers. A very few such cells subdivided into strands point of liber diameter as seen in a macera­ of S cells were observed. Growth rings absent, dif­ tion and averaging 50 such measurements. ferences from one portion of xyloin to another too A figure for vessels per group was not com­ minor to permit distinction as true growth rings. puted because groupings in /'. fmiiindezi- No storied structure observed. No resin-like de­ ana Fi& 3) are almost infinite; vessels posits present. in P. webbii are often as narrow as libri­ PLANTAGO MADEBENSIS Decne., Carlquist 2622 form fibei's and therefore difficult to iden­ (B8A), Ribeiro Brava, Madeira (Figs. 7-8): Ves­ tify in a transection. For <|iialitarivc data. sels 21-52 (average: 32) p. in diameter. Vessel ele­ 50 or more measurements per Feature per ments 91-178 (average: 141) ^ long. Vessels in collection were employed. Tabular sum­ radial chains, but many solitary. Perforation plates marization, as used in ray earlier papers in simple. Lateral wall pitting alternate, pits circu­ wood anatomy, is unfortunately not feasi­ lar, about 4 u. in diameter. Libriform fibers 135— 259 (average: 224) u. in length; diameter at widest. ble because so few species are involved. point 15-26 (average: 21) p. Libriform fibers not iciably wider radially than tangentially. Wall Anatomical descriptions. PLANTAGO thickness of libriform fibers about 2.."> p.. A few CENS Poir., Carlquist 2436 (RSA), Barranco do rays present in outer portions of larger wood sam­ San Andres, Tcnerifo, Canary Islands (Figs. 5-6): ples. Rays consist of erect and square cells. Pibri Vessels 18-22 (average: 87) u la diameter. Vessel form cells, some subdivided into strands of 2, ad­ its 98-302 (»ver;i•_'.•: 171) JJ. long. Vessels in jacent to vessels; these seem clearly identifiable radial Chains, s..!it;«r_\ vessels also common. Per as parenchyma cells. Growth rings present: early foration plates simple. Pita On vessel walls about wood distinguished by slightly wider vessels and 4 JA in diameter. Pits circular, alternate, a very libriform fibers. Some libriform fibers storied. \n few elliptical pita also observed. Libriform fibers deposits of resin-like materials observed. 132-311 (average: -i'i p in length; diameter at widest point 9-27 (average: 22) p.. Libriform fi­ PLANTAGO PBINCEPS Cham., collected in 1965 by bers often slightly wider radially than tangentially. .Melvern Tessene (MICH), Aina Ho Valley, Koolau Wall thickness of libriform fibers about 3 p. Bays Gap, Maui, Hawaiian Islands (Figs. 1-2, 13). Ves­ if, or a few remnants of exceptionally wide sel diameter 31-60 (average: 38) p.. Vessel ele- primary rays present U a very few rays in; 151—291 (average: 246) n long. Vessels mostly solitary, but a few in short radial chains. tardily in larger stems. A few fibriform cells pres­ Perforation plates simple. Pits on lateral walls of ent adjacent to vessels may perhaps be termed vessels alternate, pits circular, about i ji in diam axial parenchyma cells by virtue ft which et&c, Libriform fibers 110-251 (average: 187) p. Save wider apertures tlian typical for lit .ri form in length; diameter at widest point L1-2S (aver Qrowtb rings present; early wood charac­ age: L' p. Libriform fibers not appreciably wiiler terized by wider vessels and wider- libriform libers. radially than tangentially. Wall thickness of fibers which are also slightly thinner than those "f late 3 i u. Bays absent, but beginning of ray forma wood. A few or some libriform fibers storied. No tion suggested by occurrence of shorter fibers in deposita of resin-like compounds observed. what would be interfascicular areas. Growth rings PLANTAGO ABBOBESCENS, Oarlqnist 2725. from as such not present, although a few narrower rings Santa Cm* de I.a I'aima. I .a Pnhna, Canary Is­ of thin walled fibers present (bottom of Fig. 1). lands, similar in qiiantitat res to the above, Some libriform fibers storied. No deposita of resin- has the following : Or quantitative char­ like materials present. ities: rassel diameter, 33 u: vessel-element PLANTAGO WEBBII Barn., Carlquist 2494 (RSA), length, 17.! u; libriform-fiber lengtS 208 u: libri­ Caldera of El Teide, Tenerife, Canary Islands form-fiber width, 29 p.. (Figs. 9-11): Vessel diameter 7-30 (average: 17) PLANTAGO vm D \N'A Bert., F. G. Meyer l«. Vessel elements 72—170 (average: 128) p. long. 9645 (US), Masatierra, Juan Fernandez Islands Vessels in short radial chains or solitary. A few (Figs. 1, 2, 12): Vessels 21-5-1 average: 40) n viTV narrow vessels without perforation plates, and in dinim • ; 9 202-307 (avei therefore vascular tracheids, but nearly all vessels 264) n long. Vessels in radial chains of Indefinite with perforation plates. Perforation plates all sim- length, a few solitary vessels also present. Per­ ple. Pits on lateral walls of vessels alternate, cir­ foration plates simple. Pita en vessel walls alter­ cular, about 4 ^ in diameter. Libriform fibers 110- nate, about 5 p, in diameter, circular. Libriform 219 (average: 187) p, in length; diameter af wid­ fibers 271—607 (average: 445) p in length; est point 8-16 (average: 13) p. Wall thickness of eter at widest point 15-40 (average: 29) p. Ldbri- libriform fibers about 2 p. Libriform fibers not form fibers mostly much wider radially than tan­ appreciably wider radially than tangentially. Bays gentially. Wall thickness of Hbriform fibers about absent in center of stem; present in outer portions of larger stems. Ray cells square to erect. A few 3 p.. Rays absent. A few fihrifonn cells present ad­ 356 BULLETIN OP THE TORREY BOTANICAL CLUB

Figs. 14. Sections of wood of Pacific species of Plantago.—Fig. 1-2. Plantago princeps, coll. M. Tessene. Pig. l. Transection, Slowing zone of thin-walled libers below.—Fig. 2. Tangential section. Some of the libriform fibers show :i storied condition.—Figs. 3-4. Plantago fernandeziana, F. Meyer 9645. — Fig. 8. Transection. Vessels are arranged in conspicuous radial chains.—Fig. 4. Tangential sec­ tion. Tracheary elements are notably long for Plantago. The micrometer scale above Fig. 1 was photo­ graph e same scale as Figs. I 1": it allows L35 i subdivided into intervals of 10 u each. 1970] CABLQUIST: WOOD ANATOMY OF PLANTAGO 357

Pigs. B B. Beetions of wood of Macaronesian species of FtaxtOffO.—Figs. 6 6. Flantago arborescens, Caxlquisl 2436.—Fig. 5. Transection. A growth ring is visible near the top of the photograph.—Fig. 6. Tangential section. A storied pattern may be seen in some of the lihriform fibers.—Figs. 7-8. Plantago », Carlquist 2622.—Fig. 7. Transection. Most vessels are solitary.—Fig. 8. Tangential section. Beginning of ray formation may be seen near the center of the photograph. Scale of magnification above Pig. l applies to Figs. 5-8. 358 lin.LETIN OF THE TORREY BOTANICAL CLUB l VOL. •••'.

. 9-13. Wood sections of Plantago.—Figs. 9-11. Plantago webbU, Carlqnist ~2±9i.—Fit'. 9. Tran- ion. Several growth rings may be seen; deposits of unidentified materials appear dark.—Fig. 10. Tangential section. Note narrowness of vessels.—Fig. 11. Portion of tangential section at higher magni­ fication, showing a ray, left.—Fig. 12. Plantago /'. -niaicl' ~iana, Meyer 9645. Portion of tangential sec­ tion. Compare cell size to thai of P. webbii. Fig. 11.—Fig. 13. Plantago princeps, coll. M. Tessene. Tran­ section, showing margin of pith at left. Interfascicular areas are suddenly converted to fibers during commencement of secondary growth.—Figs. 9-10. Magnification shown by scale above Fig. 1.—Figs. 11- ile of magnification shown by scale above Kg. 13. I!l70j CARLQUIST: WOOD ANATOMY OF PLANTAGO 359

(Uniform veils present adjacent to vessels may per­ mesomorphy, and the reverse applies to haps be termed axial parenchyma cells by virtue xeromorphy. of their pits, which have wider apertures than typical libriform fibers. Growth rings present, char­ The data of the present study do not acterized by wider vessels and libriform fibers in answer whether the longer vessel elements early wood. A few libriform fibers storied. Deposits of P. fernandeziana (Figs. 4, 12) and P. of resin-like materials present in many cells, par­ princeps (Fig. 2) are the product of paedo- ticularly in central portions of stems. morphosis, although that explanation seems very likely (Carlquist, 1962). If length of Ecological correlations. The nature of fusiform cambial initials decreases in woods vessel-element dimensions with relation to showing juvenilism, then shorter length in ecology was examined for Asteraeeae by the Macaronesian species might be ex­ the writer (1966). Vessel elements are pected ; however, study of radial sections of shorter and wider in xeromorphie species these species suggests that they begin with in this family. \Yh<-re a taxonomic group shorter elements than do the woods of P. immigrates to an island and adapts from femancU nana and P. princeps. I feel that drier to wetter conditions, increase in ves­ increase, phylogenetically, of length of sel-element diameter and length occurs fusiform cambial initials may be possible, during this evolution. This is true in unre­ at least within certain limits; this possi­ lated groups of dicotyledons: Campanula- bility was suggested earlier (Carlquist, ceae tribe Lobelioideae (Carlquist, 1970b), 1966). Scaevola of the Goodeniaceae (Carlquist, 1970c), Euphorbia (1970d) and Echium of The accumulation of the resin-like mate­ the Boraginaceae (1970e). These trends rials in wood of P. webbii may be a response might, also be expected, therefore, in Plan­ to the xeric conditions in which it grows, tago. In fact, this proves to be true: for such accumulations seem most common Vessel Vessel-element in woods of desert plants, stem suceulents diameter length excepted. Because there is such great eco­ Species (average, n) (average, p,) logical similarity between the habitat of P. webbii 17 128 P. fernandeziana and that of P. princeps, P. maderensis 32 141 the curious difference in vessel grouping is P. arborescens 33 173 difficult to explain. This may be evidence P. princeps 38 246 P. fernand-eziana 40 264 that these two species are not closely re­ lated but independent derivatives from Length of libriform fibers parallels this other plantagos. Too often, the similarities trend. One would expect this, for length in growth form are taken as evidence of of fusiform cambial initials governs length phylogenetic relationship, and insular ro­ of both libriform fibers and vessel elements. sette-trees and rosette shrubs tend to be One might not have expected vessel diam­ misunderstood on this account. eter to parallel length of tracheary ele­ ments so perfectly, but evidently these re­ Raylessness. All woods of Plantago spond to the same ecological factors. The begin with a rayless condition as soon as distinctions in the table above may not production of secondary xylem commences. seem very great, but the range of differ­ As shown in Fig. 13, interfascicular areas ences is perhaps compressed owing to the a iv converted to fibers rather suddenly. fact that in this specialized genus length Rays, if produced in species with rayless of tracheary elements is close to the mini­ woods, occur only in outer portions of the mum for dicotyledons. secondary xylem, as noted by Barghoorn Diameter of vessels is wider in early (1941) for Geranium tridens. Such rays wood of ring-porous dicotyledons at large, are shown here for P. webbii (Fig. 11) ; so that one would naturally expect wider rhey appear to be produced in the Plantago vessels to be an indication of mesomorphy species with more extensive accumulation and, therefore, dicotyledons of mesic situa­ of secondary xylem. The fact that Plantago tions to have, on the average, wider ves­ is both woodier on islands and also rayless sels than those of xeric regions. This ap­ is interesting; one might not think this pears to be true, on the basis of Asteraeeae unusual except that the pattern is repeated (Carlquist, 1966). In all likelihood, wider in other insular dicotyledons. Other di­ vessels are not a phenotypic modification cotyledonous groups that are basically but a genetically-controlled adaptation to herbaceous and show both increased woodi- 360 BULLETIN OP THE TORREY BOTANICAL CLUB [VOL. 1»7 nrss am! raylessness in the Hawaiian flora the year-long growing season might con­ include Geranium. (Barghoorn, 1941), Lysi- form to this situation, and virtually all of imirlii't and Viola (Carlquist, unpublished). the known rayless genera might be said to In New Zealand, Alseuosmia is rayless have these specifications. (Barghoorn. 1941; Paliwal and Strivas- (3) Groups in which fusiform cambial lava. 1969). Many dicotyledons in insular initials are relatively short. These are, of situations are not actually rayless but show course, many of the same groups that are a condition very similar morphologically: herbaceous, for herbaceous groups do tend predominance of erect ray cells, often with to have, in general, specialized wood fea- I absence of procumbent cells (Carlquist, tores of which short fusiform cambial li'Tub, 1970c, 1970d). initials is one. In a group with kmg fusi­ A number of rayless woods have now form eambial initials, the disparity between ( been reported. Some of these are listed by the length of fusiform initials and that of Barghoorn (1941). Others have been given ray initials is great; the difference musl by Boureau (1957). From my own experi- !)•- little or none for raylessness to occur. enee, I can add Jaeobima cirnea (Acan- (4) Groups in which a form of juvenil- thaceae) and Leptodachflon californicum ism or paedomorphosis occurs. As specified • I'olemoniaeeae). The number of rayless in the .in point, a group in which fusi­ woods now known is great enough so that form and ray initials are similar in length commentary concerning raylessness seems is needed for the occurrence of a rayless warranted. Although no class of dicotyle­ wood. This situation would be altered, how­ dons is predominantly rayless and. indeed, ever, and raylessness lost, if subdivision of only a tiny proportion of dicotyledonous ray initials occurred. This actually hap­ woods are rayless, raylessness appears to pens in some rayless woods that ultimately occur in the following classes of dicotyle­ form rays, e.g., Plantogo m hbii or Gera­ dons : nium Iridi ns. One may, of course, ask wh\ formation of rays should be forestalled or (1) Groups in which cambial activity delayed. Do plants of an herbaceous an is limited or finite. One group of examples try or with relatively little secondary xylem is offered by genera siu-h as I'luntago, aeciimulation have normal xylem function Viola, etc., which are herbaceous with very without rays, and is ray parenchyma neces­ little secondary xylem except where special sary for the normal function of large conditions, as the uniform climate of cer­ masses of secondary xylem? There is pre­ tain sub-tropical islands, permit a slight sumably a selective value in the preset increase in production of secondary xylem. and structure of rays. Anomalous secondary growth is probably The above considerations do not totally , result, in at least sum.' --roups, of loss explain ra 3, obviously. For example. of normal cambial activity during evolu­ no explanation is offered why the cambium tion toward an herbaceous mode of struc­ should add thick-walled cells (libriform ture. If a single cambium can no longer fibers) to interfascicular areas rather than produce more than a finite amount of vascu­ thin-walled cells (parenchyma cells) as lar tissue, then successive cambia can pro­ secondary growth begins (Fig. 13). The duce a larger stem or root diameter. That nature of phloem in species lacking rays anomalous woods are often rayless was also has been neglected and needs inv noted by Barghoorn (1941) and has also gation. reaffirmed in the ease of Bougain- rilha by Esau and Cheadk .'1969). Literature Cited BARGHOORN, E. S. 1941. The ontogenetic develop­ (2) Groups in which woodiness is in ment and phylogenetic specialization of rays the process of increase, rather than de­ in the xylem of dicotyledons. III. The elim­ crease. The starting point is an herbaceous ination of rays. Bull. Torrey Bot. Club 68: condition, not a woody one. This situation 317-325. BOUREAU, E. 1957. Anatomie vegetale. Vol. III. is represented on insular and tropical or subtropical areas where uniform climatic Presses Universitaires de France, Paris. BURCHARD, O. 1929. Beitrage zur Okologie und conditions permit a more nearly indefinite Biologie der Kanarenpflanzen. Bibliotheca growth season and. therefore, a more nearly Bot. 9:81-262. indefinite accumulation of secondary xylem. CARLQUIST, 8. 1962. A theory of paedomorphosis in A group of dicotyledons which immigrates dicotyledonous woods. Phytomorphology to -inch uniform situations and adapts to 12:30^5. 1970] CABLQUIST: WOOD AN\T..\IY OK PLANTAGO 361

. 1965. Island life. Natural History LEMS, K. 1960. Floristic botany of the Canary Is­ Press, New York. lands. Sarracenia 5:1-94. . 1966. Wood anatomy of Compositae: MELLISS, J. C. 1875. St. Helena. L. Reeve & Co., A summary, with comments on factors con­ London. trolling wood evolution. Aliso 6:25-44. METCALFE, C. R., and L. CHALK. 1950. Anatomy . 1970a. Hawaiian natural history. of the dicotyledons. Vol. II. Oxford Uni­ Natural History Press, New York. versity Press, London. . 1970b. Wood anatomy of Lobeli- PALIWAL, G. S., and L. M. SRIVASTAVA. 1969. The oideae (Campanulaceae). Biotropica 1:47- cambium of Alseuosmia. Phytomorphology 72. 19:5-8. • . 1970c. Wood anatomy in Goodenia- ceae and the problem of insular woodiness. ROCK, J. F. 1920. The genus Flantago in . Ann. Missouri Bot. Gard. 56:358-390. Amer. J. Bot. 7:195-210. . 1970d. Wood anatomy of Hawaiian. Si HKKCK, H. 1907. Beitrage zur Kenntnis der i Macaronesian, and other species of Euphor­ Vegetation der Canarischen Inseln. In C bia. Biol. J. Linnaean Soc. London (in Chun, ed., Wissenschaftliehe Ergebnisse der press). Deutsehen Tiefsee-Expedition auf dem . 1970e. Wood anatomy of Echium Dampfer "Valdivia" 1898-1899, 2(1:2): (Boraginaceae). Aliso 7:183-199. 1-406. ESAU, KATHERINE, and V. I. CHEADLE. 1969. Sec­ SKOTTSBEEG, C. 1952. The vegetation of the Juan ondary growth in Bougainvillea. Ann. Bot. Fernandez Islands. Nat. Hist. Juan Fer­ 33:807-819. nandez & Easter Is. 2:793-960.