Rhizophores in Rhizophora Mangle L: an Alternative Interpretation of So-called " Aerial Roots "

Rhizophores in Rhizophora mangle L: an alternative interpretation of so-called "aerial roots" ABSTRACT Rhizophora mangle L., one of the most common mangrove species, has an aerial structure system that gives it stability in permanently swampy soils. In fact, these structures, known as " aerial roots " or " stilt roots " , have proven to be peculiar branches with positive geotropism, which form a large number of roots when in contact with swampy soils. These organs have a sympodial branching system, wide pith, slightly thickened cortex, collateral vascular bundles, polyarch stele and endarch protoxylem, as in the stem, and a periderm produced by a phellogen at the apex similar to a root cap. They also have the same type of trichosclereid that occurs in the stem, with negative geotropism, unlike true Rhizophora roots, which do not form trichosclereids at all. On the other hand, these branches do not form leaves and in this respect they are similar to roots. These peculiar branches are rhizophores or special root-bearing branches, analogous to those found in Lepidodendrales and other Carboniferous tree ferns that grew in swampy soils.


INTRODUCTION
One of the few tree species of the Brazilian mangrove is Rhizophora mangle, belonging to a widespread genus in the Americas, Africa, Asia, Madagascar and Australia (Juncosa and Tomlinson 1988a).According to Juncosa and Tomlinson (1988b), the generic epithet means " root-bearer".However, according to Plumier (1703), the name Rhizophora was attributed by G. Pisone to the fact that in the propagule of the viviparous plant " the radicle is located at the extremity of an axis, the rhizophore".In fact, Pisone considered the rhizophore to be the exposed hypocotyl of the viviparous propagule itself.
One of the most striking features of this species is the presence of structures that expand its supporting base.These structures are defi ned as aerial roots by most authors, including Warming (1883), Hou (1958), Gill and Tomlinson (1969, 1971a, b, 1977), Sporne (1974), Chapman (1976), Hallé et al. (1978), Ellmore et al. (1983), Tomlinson (1986), Juncosa and Tomlinson (1988a, b), Mauseth (1988), Huang and Huang (1990) and Raven et al. (1992).Some of these authors have carried out extensive anatomical studies on these structures (Gill and Tomlinson 1971a, Chapman 1976, Ellmore et al. 1983), and concluded that they are roots, although they mention that they have detected stem-like characteristics in these organs.They also mention a strong characteristic of roots, which is the presence of a root cap.Pitot (1951Pitot ( , 1958)), in studies on Rhizophora racemosa, not only always placed the term " stilt root" in inverted commas, (" racine échasse"), but also referred to these structures as rhizophores, and advocated that they are part stem and part root.According to Pitot (1951Pitot ( , 1958)), the rhizophore grows with a stem-like structure (sometimes as much as several meters) until it reaches the swampy soil, the apex of the rhizophore transforming itself into a root upon contact with the water.According to this author, the submerged portion is, therefore, a root, and all the aerial part, a stem.
On the other hand, Huang and Huang (1990), working with several mangrove species, refer to the fact that " the structure of the aerial root in Rhizophora mangle resembles that of the stem", although they do not explicitly refer to the structure as a stem.
The presence of H-trichosclereids in the cortex was observed, in both the stem and root of Rhizophora mangle, by Gill and Tomlinson (1971a, b) and in the stem and aerial root, by Warming (1883) and Karsten (1891).The last two authors demonstrate that there is a system of brachiform cells in the submerged root cortex, with a special thickening in the cell walls, which prevents the collapse of the cell due to the large air spaces within it.Referring to these cells observed by Warming (1883), Gill and Tomlinson (1971a) propose that this was a technical flaw, once Warming's observations were made on pickled specimens.Gill and Tomlinson (1971a, p. 63) emphasize that, as their laboratory was directly in front of the mangrove, they were working with recently collected material, and were able to affi rm that the root does have H-trichosclereids.
The present study offers an alternative interpretation for the aerial structure system that provides stability to Rhizophora mangle in swampy soils, and seeks to understand this structure by comparing it with rhizophores, which are root-bearing organs of the Carboniferous Lepidodendron (Stewart 1983), a plant which also grows in swampy soils.In relation to Lepidodendrales, Stewart (1983, p. 103) writes that: " these plants have a main axis that grows and branches at both ends.The branches of the aerial part form a three-dimensional system of dichoto-mous or pseudomonopodial branches, with spirally arranged leaves and terminal cones.The basal end also branches dichotomously to form the anchoring and water-absorption system, which is comprised of rhizophores bearing spirally-arranged roots".
Rhizophores have also been described in Selaginella as a sui generis organ (Nägeli and Leitgeb 1868 apud Jernstedt and Mansfi eld 1985, Goebel 1905, Jernstedt et al. 1994).Working with Dioscoreaceae, Goebel (1905) refers to the " relations between the thickened organ of Dioscorea and the rootforming organ of Selaginella".According to Goebel (1905), this organ is neither root nor stem, but a sui generis organ, half way between stem and root.Ogura (1938), also working with Dioscoreaceae, calls these thickened organs rhizophores.

MATERIALS AND METHODS
The plant material used was collected from the mangrove on the Rio-Santos highway, at Km 197, in the Municipal District of Bertioga, next to the Guaratuba River .(Menezes s.n.,SPF 124.080) on 02/25/1997.
Free-hand cross-sections were taken from the stem, root and rhizophore of Rhizophora mangle L., and stained with astra blue and fuchsin, according to the method described by Roeser (1962).The sections, mounted in 66% glycerine, were photographed using a VANOX model Olympus photomicroscope.

RESULTS
Rhizophora mangle almost always grows in swampy soils and has a positive geotropic branching system (Fig. 1 and 2) which is the result of the development of secondary rhizophores, that emerge from the erect stem and grow towards the soil, branching sympodially (exactly like the aerial branch with leaves), and expanding the supporting base of the plant.The hypocotyl of the seedling (Fig. 3) comprises the primary rhizophore, which germinates on the parent plant and generally buries itself when it falls from the tree (Fig. 4).When the secondary rhi-An Acad Bras Cienc (2006)  78 (2) zophore, protected by an apical periderm (Fig. 5), reaches the water in its positive geotropic growth, it forms roots at its extremities (Fig. 6) or around it (Fig. 7).
A cross-section of the aerial axis of a secondary rhizophore (Fig. 8-9) shows a wide pith and many bundles, in which the protoxylem is in an endarch position, surrounded by fi ber strands throughout the perimedullary region shown by the arrows in Fig. 8.The metaxylem is external to the protoxylem, where cambium, secondary phloem and secondary xylem can also be seen.The endodermis is external to the vascular system.Even when secondary growth has been fully established (Fig. 10), fi ber strands can be seen around the primary xylem.There are H-trichosclereids both in the cortex and in the pith, as shown in longitudinal section in Fig. 11.The root apex in longitudinal section (Fig. 12) shows a root cap formed by layers which separate easily, produced by a calyptrogen.In the longitudinal section of the apex of the secondary rhizophore (Fig. 13), a phellogen can be seen, which produces a protective periderm.In a submedian longitudinal section of the secondary rhizophore (Fig. 14) procambial strands can be seen, and in cross-section (Fig. 15), appearing inside the endodermis.These procambial strands present protoxylem in their interior at the level of Fig. 16.Closer to the apex (Fig. 17-18) the primary xylem forms collateral bundles with primary phloem strands.The cambium is formed in the vascular bundle and inside the interfascicular phloem.The bundles show endarch differentiation of the primary xylem.There are numerous thickened structures that correspond to H-trichosclereid arms, which are clearly visible in Fig. 11.At the level corresponding to Fig. 19, the cells are already fully lignifi ed around the primary xylem, forming fi ber strands.
In the apex of the aerial stem with negative geotropism (Fig. 20-21), a band of primary phloem can be seen, originating from the pericycle, and interspersed with bundles containing protoxylem and metaxylem, as well as primary phloem.At the level corresponding to Fig. 21, the presence of primary xylem, i.e. proto-and metaxylem, can be seen only inside the secondary xylem, formed by the cambium of procambial origin.
The roots are adventitious, and have a relatively wide pith (Fig. 22) with exarch protoxylem (Fig. 23-24).The thickened root cortex (Fig. 22) has brachiform cells with lignifi ed secondary walls (Fig. 25) in certain regions of the cell.In a secondary structure (Fig. 24), parenchymatous rays formed by the cambium of pericyclic origin also can be seen in the root.

DISCUSSION
Although the branching system that supports Rhizophora mangle plants in swampy soils has always been referred to as being composed of " aerial roots" or " stilt roots", its true identity can only be resolved by means of anatomical studies.Pitot (1958) identifi ed an inversion in the vascular tissues of these organs in Rhizophora racemosa G.F.W. Meyer, which he linked to the position of the protoxylem, typically exarch in roots and endarch in these aerial branches, as in stems.He also found another stem characteristic in this organ, namely, the presence of collateral bundles.However, he identifi ed the existence of phloem strands interspersed between these bundles in a radial position, which led him to identify these structures as roots.
When only one adventitious root is formed at the apex (as shown in Fig. 6 with two roots at the apex), it simulates a hypocotyl-radicle axis, as a result of tissue continuity, unlike subapical adventitious roots.
On the other hand, an important observation by Pitot (1958) lead the author to admit another possible explanation for the appearance of roots at the apex of the rhizophore.According to Pitot, " Rhizophora racemosa, from a morphological point of view, presents two types of root: adventitious fasciculated roots, essentially aquatic, which consist of a fairly dense involucrum around a 'pivot', forming upon contact with the water; and stabilizing roots, originating from the 'pivot' which sink into the mud and transform themselves into roots, upon contact with the swampy soil" (Pitot 1958(Pitot , p. 1112)).Pitot (1958Pitot ( , p. 1118) also states that: " the transformation of an aerial organ into an underground organ is observed; from endarch bundle into alternate exarch bundle, with a rhizophore root structure.This zone or region of transformation corresponds to the submerged region".This observation by Pitot (1958) corresponds to what is normally seen in tissue culture, that is, the transformation of organs with the An Acad Bras Cienc (2006)  78 (2) use of hormones.However, as Pitot (1958) demonstrates in his work with Rhizophora racemosa, this phenomenon also occurs in nature.The same pro-cess may also occur in R. mangle, but I believe that a single root primordium may also form at the apex of the rhizophore, similar to that seen in this work, with An Acad Bras Cienc (2006)  78 (2)  two adventitious roots at the apex of the rhizophore, resulting from two root primordia.
While in this study, only regions with stem characteristics are considered rhizophore, Pitot (1958) considered the whole structure to be a rhizophore (with both stem and root regions), hence, he referred to it as an " intermediate organ".However, Pitot did not explain his reasons for labeling the organ a rhizophore.
The presence of interspersed or radial primary phloem strands between the bundles was indicated by Pitot (1958) 1976) made a series of important observations on the adventitious origin of these rhizophores, which they refer to as " aerial roots".In other words, unlike roots (except for the radicle), they are not endogenous in origin.According to these authors, these branches are clearly adventitious because the primary xylem does not have a vascular connection with the primary xylem of the stem that originated it.Ac-cording to Chapman (1976), " there are no prior root primordia, neither can continuity with the primary xylem" be observed between the root and its generator axis.Therefore, like the stem, they are exogenous in origin.Gill and Tomlinson (1969) demonstrate that while all roots have a monopodial branching system, the " aerial roots" of Rhizophora have a sympodial branching system (a characteristic shared with stems) which is also demonstrated in this work.
Gill and Tomlinson (1971a, b) also observe that earlier authors questioned the root nature of these branches due to the presence of trichosclereids.They believe mainly owing to the presence of wide pith and polyarch stele, the structure is also rather different from the characteristic structure of other dicotyledonous roots.Gill and Tomlinson (1971a, b) also mention that in these organs, the protoxylem is on the inner side of the metaxylem and they believe that this condition led to confusion among earlier anatomists, as to whether or not it was a root.The authors conclude their study by emphasizing features which are not found in dicotyledonous roots, and which are considered exceptions in the roots of Rhizophora mangle: polyarch stele, wide pith, the collateral position of the vascular tissues, and endarch protoxylem.In a later study, Gill and Tomlinson (1977) also mention that these " aerial roots" originate in the trunk, or in other " aerial roots", sympodially, but never in underground (or submerged) roots.
The results presented in this study indicate that the aerial branch system of Rhizophora mangle, with positive geotropism, is, in fact, a rhizophorous system which is very similar to a stem system with negative geotropism, as perfectly demonstrated in Table I of the present work.All the characteristics which are considered exceptions in roots, defi ne the rhizophore as a stem system: to these characteristics, one can add the presence of H-trichosclereids identical to those found in rhizophores and negative geotropic stems, and distinct from Warming root cells, with phi-thickenings (Haas et al. 1976).Another major difference between the rhizophore and the root, as demonstrated in this work, is that while in the true root of Rhizophora mangle there is a root cap, formed by the calyptrogen, in the rhizophore, the apical protection is a periderm formed by a phellogen.Above all, attention is drawn to the fact that the root has a subapical meristem similar to all mono-and dicotyledonous roots, while the rhizophore does not present the same type of subapical region.It grows as an extension of the apex, in which vascular tissues are formed from the procambial strands.I agree almost entirely with the statement of Pitot (1958Pitot ( , p. 1136) ) that: " This study of the anatomic transformation of the rhizophore leads us to conclude that the internal structure of the rhizophore known as the " stilt root" of Rhizophora racemosa does not correspond to its external mor-phology, which is that of an aerial root with positive geotropism, and a root cap at the extremity.Due to its anatomical structure, the rhizophore is not a root in the exact sense".The only point with which I disagree is his affi rmation that there is a root cap at the extremity of the rhizophore in fact, a periderm.However, the absence of leaves and nodes, which are characteristics of roots, supports the idea of a rhizophore as an intermediate organ between root and stem (Goebel 1905).
An analogy with Carboniferous Lepidodendrales (Stewart 1983, Gifford andFoster 1988) clarifi es the relationship between the rhizophore and the stem.An interesting point is the similarity of the Lepidodendron reconstruction with my Fig. 2, a sympodial branching with leaves and the basal system of rhizophores, also with sympodial branching.Despite the diffi culties involved in studying fossilized plants, Stewart (1983) reconstituted these basal organs in Lepidodendron, demonstrating that they also have a cauline structure.Ac-cording to Stewart (1983), Stigmarian systems are comprised of roots linked to a root-bearing axis, the rhizophores.These rhizophores are known as anchoring systems, and form the absorption system of Lepidodendrales.
I have already identifi ed rhizophores in many

Root cap --
Reconstruction of Lepidodendron sp.(Stewart 1983) other angiosperm families.When I fi rst observed a second cauline system, with positive geotropism in the Asteraceae genus Vernonia (Menezes 1975, Menezes et al. 1979) I assumed it to be a mutation and, in an analogy with Selaginella (Selaginel-laceae), I chose the term rhizophore.In choosing this label, I took into account the fact that angiosperms have roots, stem and leaves, like Pteridophyte, and so rhizophore seemed an appropriate choice.Later, while studying members of Dioscoreaceae (Rocha and Menezes 1997) and Smilacaceae (Andreata and Menezes 1999), a comparative analysis of these two families led the discovery that the tuberized structures are indeed rhizophores, as in Vernonia.Our work on Dioscoreaceae (Rocha and Menezes 1997) led us to a study by Goebel (1905), who lamented the fact that up until then, morphologists had failed to notice the existence of what he referred to as " an intermediate organ between root and stem in the Dioscoreaceae, analogous to Selaginella's root-bearing organ".It was only after studies on Dioscorea (Rocha and Menezes 1997) and Smilax (Andreata and Menezes 1999) had been published, that we became aware of Ogura's study (1938), which contained the same considerations on Dioscoreaceae as ours, i.e., analogies with Selaginella and Lepidodendrales, and a reference to Goebel.I consider the possibility of the rhizophore, by evolution, gives rise to the rhizome, rather than the latter originating from an aerial stem, as is currently assumed.Although the rhizophores in Rhizophora are above the surface, as in Lepidodendrales, in other angiosperms this organ is normally found under the ground.Tomlinson (1962, p. 211), in his studies on the phylogeny of Scitamineae, makes an interesting statement.Referring to the species Phenakospermum, he writes: " It is one of the dogmas of elementary botany that the rhizome is morphologically equivalent to a stem modifi ed as a horizontal, storage and propagating organ.Can this long-accepted idea be challenged?The rhizome, or its equivalent, already exists in primitive monocotyledons.This might suggest that it is an organ sui generis and not homologous with the aerial stem.One is tempted to compare the rhizome in the Scitamineae with Stigmarian axis of Lepidodendron and the rhizophores of Selaginella similar organs, the morphology of which is unexplained".
I propose here that Rhizophora mangle has a rhizophore system.Juncosa and Tomlinson (1988b) state that the genus name Rhizophora means " rootbearer".However, I believe that Rhizophora should mean " one that has rhizophores" and that rhizophore is a " root bearing" branch.
At fi rst, I did not believe that Pisone's rhizophore (according to Plumier 1703) bore any relation to the rhizophore I was describing.However, it is the very fi rst rizophore of Rhizophora mangle, hence it is referred to in this paper as the primary rhizophore (Rh1), i.e., with well-developed hypocotyl, unlike the secondary rhizophores (Rh2), which appear later in R. mangle from adventitious buds.
According to Chapman (1976), no primary root appears to develop in Rhizophora mangle.In a study which is not yet published, I intend to demonstrate that this is, in fact, true.

ACKNOWLEDGMENTS
The author thanks Vanessa de Aquino Cardoso and Delmira da Costa Silva for the cross and longitudinal sections of rhizophores and plate preparation; Antonio Salatino, Mary Gregory and Daniela Zappi for English revision; D. Zappi, Simon Mayo, David John Nicholas Hind, from RBG Kew, and Orbelia Robinson, for access to essential bibliography; Norberto Palacios who helped with the text and plates; Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), for the support provided (Process 93/2444-8 and 2005/54439-7) and Conselho Nacional de Desenvolvimento Científi co e Tecnológico (CNPq) for research grants.Palavras-chave: Rhizophora mangle L, rizóforo, " raízes aéreas", " raízes suporte".

Dedicated with much admiration to Dr. David F. Cutler, as a homage to him in his retirement as Head of Kew's Anatomy Section, for his important contributions to our understanding of Plant Anatomy, and for his really excellent and warm welcome to all of us, including myself, from diffent parts of the world, who benefited from his extensive knowledge to widen our own experience of the subject.
Figs. 1-2 -Rhizophora mangle in Porto de Galinhas (Pe, Brazil).Fig. 1 -Entire plant with the secondary rhizophores (Rh2) highlighted, at the base.Fig. 2 -Stem branch with leaves, with the same sympodial branching system as that seen in secondary rhizophores.(Photo Ricardo Sacerdoti)

Figs. 23
Figs. 23-26 -Cross-sections of the roots of Rhizophora mangle.Fig. 23 -Adventitious root with a well-developed pith (Pi) and cortex (Ct) with a thickness of 3 times the diameter of the vascular cylinder.Fig. 24 -Region of the root in primary structure with exarc protoxylem (Px); white arrow indicates secondary xylem.Fig. 25 -Region of the root in secondary structure.Fig. 26 -Warming cells (W) with a special type of thickness, lignifi ed; the arrows indicate regions rich in plasmodesmata, in the contact between two cells.En -endodermis; Mx -metaxylem; Ph -phloem; R -parenchymatous ray; Xy -secondary xylem.The bars correspond respectively to: 2 mm, 100µm, 150µm and 60µm.