An Anatomical Study of the Hawaiian Fern Adenophorus sarmentosus'

KENNETH A. WILSON2 and FRED R. RICKSON:;

THE ENDEMIC HAWAIIAN FERN, Adenopborus tative means of propagation is known not only sarm entosus ( Brack.) K. A. Wilson, occurs on in A. sarmentosus, but also in the closely re­ all of the major islands, growing on moss­ lated A. haalilioanus, and should be looked for covered trees and occasionally also on mossy in the rare A. pimlatifidus. Similar reproductive rocks. As a representative of the large but poorly behavior has been reported recently in Asple­ understood fern family Grammitidaceae, A. sat­ nium plenum of Florida (Wagner, 1963). mentosus was chosen for anatomical studies in The petioles are short, less than 2 em long, order to contribute some information on the crowded on the rhizome, and bearing simple or family which may be of value in later systematic branched deciduous, reddish-brown hairs. The .studies of the group. Anatomical or morpho­ blades are pinnatifid, elliptic-lanceolate, 8-15 cm logical studies of the grammitids are rare. A long, 1-2 .5 cm wide, and narrowing gradually series of recent papers by Nozu (1958-1960) at both ends, often becoming prolonged into a presents the only anatomical investigation of long caudate apex (Fig. 1, 3) . The venation is members of this family except for a few notes pinnate, with free simple (rarely branched ) published earlier by Ogura (19 38) . veins with clavate or punctiform ends which do The plant material used was collected at W ai­ not extend to the margins of the blade ( Figs. anae Iki on the island of Oahu in 1954 by Earl 1, 3 and 4). T. Ozaki. A voucher specimen has been de­ The outline of the stem is almost a perfect posited in the herbarium of San Fernando Val­ circle in cross-section. The vascular cylinder is ley State College. The fern was killed and fixed prorostelic in the young stages and becomes in F.A.A.; sectioning was done according to solenostelic in older portions of the rhizome. standard paraffin techniques. Leaves were cleared During the transition from the prorosrele, a by the sodium hydroxide technique (Foster, parenchymatous pith develops ( Fig. 2, 1 ), 1949 ) and stained in tannic acid and ferric within which a central phloem strand differen­ chloride (W ilson, 1958 ). tiates in slightly later stages ( Figs. 2, 2 and 3 ). In the older portions of the stem the inner OBSERVATIONS phloem of the solenostele is cylindrical and sur­ The rhizome of A. sarmentosus is erect or rounds a central pith (Figs. 2, 13-17). The single ascending, short, usually less than 1 em in length, leaf traces which arise from the main stele and Qj-l mm thick, and clothed in reddish-brown, lead to the leaves are surrounded by the scleren­ concolorous, linear -lanceolate scales (Fig. 1, 8). chyma of the cortex which extends into the The roots are wiry, dark brown, and few­ petiole itself. Each leaf trace is generally asso­ branched, often bearing propagules which de­ ciated with two root traces which arise either velop into new plants (Fig. 1, 2) . This vege- below the juncture of the leaf trace with the stelar gap (Figs. 2, 2-5) or somewhat above this level (Figs. 2, 7, 29, 30). One instance was 1 This project was supported by a grant from the Nati onal Science Foundation and by a summ er re­ observed in which a leaf gap was associated with search grant to the junior auth or from the California only a single root trace which arose directly Foundation for Biochemical Research. Manuscript re­ oppos ite the gap rather than perpendicular to ceived October 19, 1964 . the plane of the gap ( Fig. 2, 21). Not all 2 Department of Biology, San Fernando Valley State roots, however, arise in association with leaf College, Northridge, California. 3 Present address : Department of Botany, Un iversity traces. Pairs of root traces were observed de­ of California, Berkeley, California. parting from the main stele in regions between

114 Anatomy of Adenophorus sarmentosus- WILSON and RICKSON 11)

FIG. 1. Adenophorus sarmentosus. 1, Habit X Y3 ; 2, habit showing vegetative reproduction by root prop­ agules, X Y3 ; 3, frond, X ~ ; 4, porti on of frond showing venati on, X 3 J;2 ; 5, cross-section of stele" of rhizome X 100; 6, cross-section of stele of root, X 100 ; 7, cross-section of stele of petiole, X 100; 8, rhizome scale, X 17; 9, cross-section of lamina, X 100 (stippled porti ons indicate cut sur faces); 10, upp er epidermal cells, X 100; 11, lower epiderm al cells, X 100; 12, epidermal hairs, X 100. successive leaf traces, showing no particular spicuous, bright-staining Casparian strips on the relation with the vascular supply to the leaf radial walls. In all of our material the stele is (Figs. 2, 13-16). The orientation of the root pulled away from the outer tissues of the stem traces is horizontal as they arise from the stele at the region of the endodermis (Fig. 1, 5 ). as well as during their course through the cortex. This separation results in the destruction of The leaf gaps do not become closed in all cases most of the endodermal cells and apparently before the departure of the next leaf trace. In is caused by the differential action of killing our material the gap was observed to close in and fixing agents on the tissues of the stele the lower portion of the rhizome (Figs. 2, uersus the surrounding cortical tissues. 10-12), while in the upper portions no such The cortex is composed of 10-12 layers of closure was observed ( Figs. 2, 19-31). cells. The cells of the innermost 5- 7 layers con­ The stele is surrounded by an endodermis sist of heavily thickened , often completely oc­ composed of small, thin-walled cells with con- cluded, sclerenchyma cells. Howe ver, patches 116 PACIFIC SCIENCE, VAI. XX, January 1966

400 lJ so SI ~ I I 1-31 , h s sarmentosus. FIG. 2. Adenop oru d root traces. on. gln. of leaf traces an

. lIed rotoplast-conta ~ning h outer rnargiin of the xylemI of slightly thinner-wa ll'yPdirectly opposite a tive areas on t e therefore, is apparent y lIs often occur, usua 2 or 3 cortical ce f The outer cylinder. The xylem' r der of xylem, the cells departing lea trace. d of uniformly large exarch. Within the cy 10 . ally two-layered d or occasion cell layers are composeII of the single-rowed of the one-Iayere d in most instances ells The ce s hI re separate II parenchyma c .. aIled internal p oem a I of parenchyma ce s epidermis are thin-w . f om one to three cells from the xylem by a ay~r cells of the central The xylem cylinder IS rf relatively uniform similar in structure t~ t a; enchyma also occurs thick Its tracheids are ch .ds are only about pith. Another layer 0 dPthe outer phloem. The . lem trac ei I size. The protoxy erage metaxylem e e- between the xylem anhloem are conspicuously . of the av d' . Y2 the diameter d in more or less isnnc- cells of the outer pf the inner phloem and ments and are groupe smaller than those 0 Anatomy of Adenophorus sarmentosus-WILSON and RICKSON 117 have a diameter of only Y2-!0 of the average four-celled, uniseriate, simple or frequently of the elements of the inner phl oem. In the branched, and with a large, redd ish-brown ter­ outermost tissue of the stele, outside the narrow minal cell ( Fig. 1,12 ) . strip of outer phloem, occur the numerous large In cross section, the cells of the upper epi­ cells of the pericycle (Fig. 1, 5 ) . dermi s are seen to have irregular vertica l pro ­ The stele of the petiole is circular in cross jections which reach below the epidermal layer section and is surrounded by flat endod ermal itself and come in contact with the spongy tis­ cells which separate it from the heavily sclerified sue of the leaf ( Fig. 1, 9) . The lower epider­ layers of the cortex. The center part of the stele mal cells rarely display such projections. The is occupied by the single xylem strand which, mesophyll is composed enti rely of irregular, in cross section, is roughly ren iform ( Fig. loosely arranged, spo ngy parench yma cells with 1, 7). The proroxylem elements are in groups large intercellular spaces. The round sari are of about six at each of two points forming the borne terminally on the veins. The sporangia adaxial lobes of the strand. The xylem is sep­ have a single row of cells at the base of the arated from the phloem by two or three layers stalk, and an annulus with 11-13 thickened of irregularly shaped, nucleated parenchyma bow cells (W ilson, 1959a). The spores are cells. The phloem area is made up of one or green, tetrahedral, trile te, without exospore, and two rows of tiny, thin-walled cells, and sur­ have an average measurement of 34 X 12 p­ round s the entire xylem core except at its (Selling, 1946). protoxylem poles. Between the phl oem and the The roots have a diarch stele with two pro­ endodermis occur the large, nucleated, thin­ roxylern group s, one at each end of the xylem walled cells of the pericycle containing plastids. tract ( Fig. 1, 6 ) . The one- or two-layered The cortex is clearly differentiated from the phl oem surrounds the xylem except at the pro­ epid ermis and is composed of six to eight layers roxylern poles, and is separated from it by a of thick-walled sclerenchyma cells. The thick ­ layer of parench yma cells. The endodermal cells ness of the cell walls of the cortical cells in­ are small, and thin- walled, and were difficult creases gradually, so that the cells in the outer­ to observe in our preparations. Thi s many­ most layer have "extremely thick walls. There celled endodermis is surrounded by the "inner is also a gradual decrease in the diameter of layer of the two layers of heavily sclerified the cells toward the periphery of the cortex. cortical cells. The outer two or three layers of The large thin-walled cells of the epidermis are cortical cells are conspicuously thinner-walled. subtended by a layer of slightly smaller thin­ walled cells containing num erous p' asrids which DISCUSSION separate the cortical sclerenchyma from the In general, the structure of Adenophorus epidermis itself. In the lateral margin al wings sarmentosus agrees with the observations of of the petiole the epidermis, containing occa­ N ozu ( 1958- 1960 ) made on species of Micro­ sional pairs of guard cells, is subtended by a polypodium (X iphopteris ), Scleroglossum, and tissue of loosely arranged parench yma cells with Grammitis. On the basis of these studies and large intercellular spaces. of general systematic surveys (Ho lttum, 1954; The adaxial laminar surface lacks stomata. Copeland, 1947), the solenostele of A. sarm en­ In face view the epidermal cells are approxi­ toms seems to be characteristic of the rhizome mately 1- 3 times as long as wide, and their of most of the Grammitidaceae, although in general outline is somewhat oblong. The longi­ some groups a dissected solenostele is not un­ tudinal walls are mod er at ely wavy (Fig. 1, common. The absence of an internal phloem 10 ) . The abaxial laminar surface is characterized in the younger portions of the rhizome of Ade­ by more irregular cells, generally with pro ­ nophorus sarmentosus is noteworthy. A similar nounced undul ate walls ( Fig. 1, 11 ). Both medullated protosrele was described in Micro­ epidermal surfaces bear scattered suberect or polypodium okttboi by N ozu. The stele of A . erect hairs. The epidermal hairs are two- to sarmentosus, however, soon becomes solen0 - 118 PACIFIC SCIENCE, Vo!' XX, January 1966 stelic, interrupted occasionally by leaf gaps. In by Wilson (1964 ) . We have undertaken this this it resembles the species of Grammitis in­ present study in order to obtain some informa­ vestigated by Nozu. The dicryosrelic vascular tion which may be of value in later systematic cylinder which he reported for Scleroglossum considerations of the grammitid ferns. Other pusillum evidently is not repeated in A . sar­ similar studies are badly needed. ment osus. However, the failure of some leaf REFERENCES · gaps to close may perhaps be interpreted as a trend towards the further dissection of the COPELAND, E. B. 1947. Genera Filicum, the solenostele. Conclusive evidence cannot be pre­ Genera of Ferns. Chronica Botanica, Waltham, Massachusetts. sented at this time to establish the evolutionary FOSTER, A. S. 1949. Practical Plant Anatomy. trend in the vascular structure of the rhizome 2nd ed. D. Van Nostrand, New York. in this family. The hypothesis that the family HOLTTUM, R. E. 1947. A revised classification consists of a reduction series seems to us to of the leprosporangiare ferns. J. Linn. Soc. be most acceptable. With this point of view, ( Bot.) 53:123-158. we would interpret the solenostele as having --- 1954. Flora of Malaya, II. Ferns of been derived from a dictyostele, although it Malaya. Government Printing Office, Singa­ seems probable that some dicryostelic grammitid pore. groups in turn have evolved secondarily from Nozu, Y. 1958-1960. Anatomical notes on solenostelic ancestors. The temptation to make Micropolypodium, Scleroglossum, and Gram­ broad generalization about evolutionary trends mitis, 1-3. J. Geobot. 7:102-107; 8: 1-5; is one to which we would rather not yield until 9:4-9. more comparative data are available. OGURA, y. 1938. Anatomie der Vegetations­ A leaf supplied by a single vascular trace organe der Pteridophyren. In : Karl Linsbauer, with a U- or V-shaped xylem in the petiole Handbuch der Pflanzenanatomie. Band 7, Abt, seems to typify members of this group. Ogura 2 (Teil 2). Berlin. ( 1938 ) designated this particular type of pet­ SELLING, O. H. 1946. Studies in Hawaiian Pol­ iolar vascular structure as the "Gramrnitis" type len Statistics, Part I. The Spores of the Ha­ and subsequent investigators have found similar waiian Pteridophytes. Bishop Mus. Spec. Pub!. patterns in other grammitids (Holttum, 1947; 37: 1-87, pIs. 1-7. Nozu, 1958-1960). The xylem strand of A. STOKEY, A. G. 1959. Polyp odium pectinatum sarmentosus adds still another species to this list. and P. plum ula-Polypodiaceae or Gram­ It is difficult to attempt to interpret the mitidaceae ? Am. Fern J. 49:142-146. phylogenetic relationship of the Grammitidaceae --- and L. R. Atkinson. 1958. The game­ on the basis of anatomical investigations of so tophyte of the Grammitidaceae. Phytomorph­ few members of the group. Numerous studies ology 8: 391-403. have been published which indicate that these WAGNER, W . H., JR. 1963. A biosy srernaric ferns represent a natural group. Other than ana­ survey of United States ferns-preliminary tomical studies of the sporophyte, evidence from abstract. Am. Fern J. 53: 1-16. the gametophyte generation (Stokey, 1959; WILSON, K. A. 1958. Ontogeny of the sporan­ Stokey and Atkinson, 1958 ) and from the gium of Pblebodium ( Polypodium) aureum. morphology of the sporangium ( W ilson, 1959a; Am. J. Bot. 45:483-491. 1959b ) strongly support the distinctness of the --- 1959a. Sporangia of the fern genera family. Broad phylogenetic conjectures would allied with Polypodium and Vittaria. Contrib. best await additional studies within the Gram­ Gray Herb. 185:97-127. mitidaceae. Systematic problems within the --- 1959b. The sporangia of three prob­ family will also require detailed studies before lematic species of Polyp odium. Am. Fern J. a satisfactory taxonomic arrangement can be 49 :147-151. proposed. Some of these problems have been --- 1964. Notes on the Hawaiian fern discussed recently for the genus Adenophorus genus Adenopboras. Am. Fern J. 54:68-70.