Living Cells in Wood 3. Overview; Functional Anatomy of the Parenchyma Network

Sherwin Carlquist

The Botanical Review

ISSN 0006-8101

Bot. Rev. DOI 10.1007/s12229-018-9198-5

1 23 Your article is protected by copyright and all rights are held exclusively by The New York Botanical Garden. This e-offprint is for personal use only and shall not be self- archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23 Author's personal copy

Bot. Rev. https://doi.org/10.1007/s12229-018-9198-5

Living Cells in Wood 3. Overview; Functional Anatomy of the Parenchyma Network

Sherwin Carlquist1,2

1 Santa Barbara Botanic Garden, 1212 Mission Canyon Road, Santa Barbara, CA 93105, USA 2 Author for Correspondence; e-mail: [email protected]

# The New York Botanical Garden 2018

Abstract The very different evolutionary pathways of conifers and angiosperms are very informative precisely because their wood anatomy is so different. New information from anatomy, comparative wood physiology, and comparative ultrastructure can be combined to provide evidence for the role of axial and ray parenchyma in the two groups. Gnetales, which are essentially conifers with vessels, have evolved parallel to angiosperms and show us the value of multiseriate rays and axial parenchyma in a vessel-bearing wood. Gnetales also force us to re-examine optimum anatomical solutions to conduction in vesselless gymnosperms. Axial parenchyma in vessel-bearing woods has diversified to take prominent roles in storage of water and carbohydrates as well as maintenance of conduction in vessels. Axial parenchyma, along with other modifications, has superseded scalariform perforation plates as a safety mechanism and permitted angiosperms to succeed in more seasonal habitats. This diversification has required connection to rays, which have concomitantly become larger and more diverse, acting as pathways for photosynthate passage and storage. Modes of growth such as rapid flushing, vernal leafing-out, drought deciduousness and support of large leaf surfaces become possible, advantaging angiosperms over conifers in various ways. Prominent tracheid-ray pitting (conifers) and axial parenchyma/ray pitting to vessels (angiosperms) are evidence of release of photosynthates into conductive cells; in angiosperms, this system has permitted vessels to survive hydrologic stresses and function in more seasonal habitats. Flow in ray and axial parenchyma cells, suggested by greater length/width ratios of component cells, is confirmed by pitting on end walls of elongate cells: pits are greater in area, more densely placed, and are often bordered. Bordered pit areas and densities on living cells, like those on tracheids and vessels, represent maximal contact areas between cells while minimizing loss of wall strength. Storage cells in rays can be distinguished from flow cells by size and shape, by fewer and smaller pits and by contents. By lacking secondary walls, the entire surfaces of phloem ray and axial phloem parenchyma become conducting areas across which sugars can be translocated. The intercontinuous network of axial parenchyma and ray parenchyma in woods is confirmed; there are no “isolated” living cells in wood when three-dimensional studies are made. Water storage in living cells is reported anatomically and also in the form of percentile quantitative data which reveal degrees and kinds of succulence in angiosperm woods, and norms for “typically woody” species. The diversity in angiosperm axial and ray parenchyma is presented as a series of probable Author's personal copy

S. Carlquist optimal solutions to diverse types of ecology, growth form, and physiology. The numerous homoplasies in these anatomical modes are seen as the informative results of natural experiments and should be considered as evidence along with experimental evidence. Elliptical shape of rays seems governed by mechanical considerations; unusually long (vertically) rays represent a tradeoff in favor of flexibility versus strength. Protracted juvenilism (paedomorphosis) features redirection of flow from horizontal to vertical by means of rays composed predominantly or wholly of upright cells, and the reasons for this anatomical strategy are sought. Protracted juvenilism, still little appreciated, occurs in a sizeable proportion of the world’s plants and is a major source of angiosperm diversification.

Keywords Flow in rays . Gnetales . Growth flushes . Juvenilism . Succulence . Xeromorphy

Introduction

As the physiology of water conduction in vessels and tracheids has become better understood, attention has turned to the roles that ray and axial parenchyma in functioning of secondary xylem of conifers and angiosperms. This wave of interest, most recently represented by the thoroughly-referenced papers by Morris et al. (2015, 2017) and Morris and Jansen (2016), seeks to clarify what living cells, including living fibers, do with regard to wood function. The approaches to study of parenchyma function must of necessity be different from those applied to sap-conducting cells of the wood. Axial parenchyma, and, indirectly, ray parenchyma are concerned in suppression and reversal of embolisms (Braun, 1984; Holbrook and Zwieniecki 1999;Holbrooketal. 2002; Johnson et al., 2012; Lens et al., 2013; Nardini et al., 2011;Salleoetal.,2004, Salleo et al., 2009; Secchi et al., 2016; Trifilo et al., 2014). These authors do not examine the role of ray parenchyma, but there is no other source for the sugars and ions than via the rays. In examining the function of axial and ray parenchyma, assembling of quantitative data has, to a large extent, been the method of choice (Morris et al. 2015, Morris et al., 2017 and literature cited therein). The present paper attempts to use new data in comparative anatomy obtained with light microscopy, ultrastructure as seen with scanning electron microscopy (SEM), and observations on growth form and habit primarily. References to experimental work in physiology are considered a vital parallel source of information, and are cited here, as they were in Carlquist (2012a). The value of comparative anatomical data is considerable in interpreting the functioning of wood, because wood anatomical diversity ultimately must be explained in terms of selection for structural features. The many parallel acquisitions of wood character states in clade after clade represent the results of natural experiments performed on immense numbers of individuals in thousands of species over many million years. There has been some appreciation of the value of anatomical data (e.g., Morris & Jansen, 2016). Such works, rich in illustration, as Moll and Janssonius (1909-1936) and Metcalfe and Chalk (1950) are cited by Morris and Jansen (2016). To those we must add such works as Greguss (1955, 1959), Meylan and Butterfield (1978), and many other important sources. Many of these are expensive and found in a relatively small number of libraries, and are not available on the internet, Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... which increasingly has become the source of references and the basis for documenta- tion in research. Simultaneously, there has been a disappearance from many colleges and universities of courses in plant anatomy, so that those able to do interpretive work in wood anatomy are fewer. Those with encyclopedic knowledge of comparative wood anatomy of large numbers of species, such as I. W. Bailey, C. R. Metcalfe, and S. J. Record, are almost non-existent today, although some a few specialists conversant with comparative wood anatomy of some families and clades do, fortunately, exist. Exper- imental studies of wood function are, by contrast, relatively recent and easy to access. The complexity and diversity of wood anatomy and the difficulty of access to the field have worked to the disadvantage of visual understanding of the function-structure continuum. The present paper is an attempt to work with microscopy and some allied data by way of supplying the visual component. This is done by presenting illustrations for a number of modes of structure in wood anatomy and attempting to show how they may be related to functional aspects. Cross-comparisons of wood of unrelated or distantly-related groups can be highly informative. Metcalfe (1983, p. 4) says, “The development of the vessel has had a profound effect on the xylem of angiosperms….it has made specialization possible in other directions: specialization of fibers [i.e, change from tracheids to libriform fibers]….and this in turn has been linked with changes in the distribution of parenchyma cells.” Now that we know that Gnetales are conifer derivatives (Bowe et al., 2000; Burleigh & Mathews, 2004), and that Gnetales have attained, parallel to angiosperms, essentially all of the important anatomical features of angiosperm wood (Carlquist, 2012b), we have a source for demonstrating quite dramatically pathways of wood evolution in vessel-bearing angiosperms and their significance. The wood of Gnetales, uniquely valuable precisely because of its independent acquisition of vessels, is virtually unmentioned in consideration of angiosperm wood evolution. Likewise, physiological studies of wood function have not included Gnetales. Key anatomical features are presented here as a way of showing probable associations of angiosperm- like parenchyma and other living cells in Gnetales with vessels. The method of anatomical cross-comparisons can be used quite productively as a way of generating hypotheses about the significance of divergent modes of structure within conifers and within angiosperms. Molecular phylogenies can show us that parallel evolution of axial parenchyma amounts and distributions, as well as kinds of ray histology, can be used to demonstrate not merely continuations of basic types, but sensitive adaptations to ecology and growth form. We tend to forget the ease with which genetic changes and ontogenetic variations can occur. For example, we tend to accept that the almost exclusive presence of uniseriate rays in conifers has a selective value. Conifers can produce multiseriate rays if the rays contain resin canals, and Gnetales are all characterized by multiseriate rays. The development of vessels has led to repurposing and diversification of parenchyma in angiosperms. We need to move beyond descriptive knowledge of living cell types in wood, although ironically, more comparative microscopical data, especially those obtained at higher magnifications, will help advance our understanding of cell function in wood. The three-dimensional nature of the network of living cells in wood has been established and studied, most notably by Zimmermann (1971) and Kedrov (2012). Understanding that in plants as diverse as Fitzroya (Cupressaceae) and Alnus (Betulaceae), as shown by Kedrov (2012), the network system prevails and no isolated Author's personal copy

S. Carlquist living cells exist is important to our interpretation of wood structure. In analyzing axial parenchyma, we are told that diffuse parenchyma is present as “isolated” cells, but in fact, all of these are part of an intercontinuous network. Transverse sections, often the primary basis for determining the nature of axial parenchyma, are misleading in this respect. In fact, the idea of an axial parenchyma cell not in contact with other living cells is highly unlikely because it would then be a kind of parasitic cell, gaining water and nutrition from water-conducting cells. Kedrov’s(2012) work goes far to promote a three-dimensional understanding of the living cell network as well as the way other cells in wood contact each other. Braun’s(1970) “Stufen” (stages) in parenchyma evolution were proposed prior to our knowledge of molecular phylogeny, and thus can at best be interpreted as physiological conditions divorced from phylogeny. Cell contents have been relatively neglected. The reasons for this lie largely in the use of xylarium specimens. For example, starch in parenchyma cells agglutinates into amorphous masses or, more commonly, is lost entirely by the rapid action of bacteria during the drying of wood samples. The use of liquid-preserved wood samples as a basis for studies should have become a standard methodology by now, but it is not. To be sure, liquid preservation of wood samples is logistically less easy than drying them, but has been routinely accomplished by a few individuals. The presence of starch and other photosynthates in living cells of woods is obviously important, as the work of Sauter (1966a, b)onstarchstorageand conversion into sugars in ray tissue of Acer tells us about the seasonal course of storage. Perhaps the most important (but least cited) paper with relation to the role of parenchyma in woods by releasing sugar into vessels (and thereby maintaining conduction through embolism prevention or even reversal) is that of Sauter et al. (1973). The accumulation, mobilization, and differential storage of photosynthates in particular parenchyma cells throughout seasons, especially those that feature marked temperature fluctuations, is not the same in all species, and is not difficult to study. Parenchyma cell size and shape are prime indicators of function, along with other indicators. The present paper takes, for the purposes of an initial assessment, the concept that cells are elongate in the direction of conduction and flow. Elongate cells with minimal contents are considered to be indicative of flow, and the non-committal “flow cell” is applied to them unless there are contrary indicators. For example, fibriform cells are not considered to be flow cells. A substantial end wall contact with another cell is essential. The existence of elongate—and non-elongate cells in rays has long been known. By applying the terms “upright,”“procumbent,” and “square,” to ray cells as seen in radial section, most authors conclude their investigations. The functional distinctions of parenchyma cells of different shapes is only beginning to be appreciated at the present time. The same considerations apply to sizes and abundance of parenchyma cells. The present essay makes use of information on size, shape, abundance, and degree of elongation as functional indicators. Anatomical literature has followed the dictum that “square cells [in rays, as seen in radial section] are morphologically equivalent to upright cells.” This seems reductive and a way of concealing information for the sake of nomenclatural simplicity. Square cells are obviously not upright cells, and the quantity of them produced in any given place by the plant can be significant. The point here is not to counter existing nomenclature but to re-explore ordinary light microscopy as a way of revealing functional significance. As Kedrov (2012) says, the entirety of a wood, not just individual parts of it, must ultimately be taken into account in explaining how it functions. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

The relative proportions of upright and procumbent cell in woods are not conveyed by the terms “heterocellular,”“homocellular,”“heterogeneous” or “homogeneous.” While convenient to use for initial assessments or for diagnostic purposes such as wood identification, we must realize that we are dealing with physiological and ecological wood anatomy. The fact that there are woods in which all ray cells are upright seems not to have been mentioned in wood literature prior to my paper on juvenilism in woods (Carlquist, 1962), and I now propose that we confront the physiological significance of prolonged juvenilism in woods. Cell wall thickness and pitting have generally been considered subordinate to cell types. Wood descriptions, if they mention these features at all, do not give dimensions for cell wall thickness, and pay little or no attention to pitting. The concept that all pits in ray cells are simple pits has been propagated if only by lack of contrary information. In fact, bordered pits on the tangential walls of ray cells can be found in a large percentage of angiosperm woods (Carlquist, 2007a). Such walls, easily visible in sectional view in radial sections, are almost never figured or explained. Tangential walls of ray cells in conifers have been given a separate terminology: strips of secondary walls between pits are considered “nodular” or “nodulated.” The point that large pits, either bordered or non-bordered, are present is thus lost in favor of a term that misses the fact that large pit membranes are present. The term “nodulated,” used by such authors as Panshin and De Zeeuw (1964), Hoadley (1990), and Roman-Jordan et al. (2016) should be abandoned in favor of descriptive language that shows that we are dealing with pitting, not some form of ornamentation superimposed on a cell wall. The same applies to the term “indentures” (Roman-Jordan et al., 2016). Tangential walls of ray cells and on the horizontal walls of strand cells in axial parenchyma can bear pits that differ with respect to size, density, border presence, and thickness of pit membranes compared with pits on other walls of these cells. These are features that govern flow of liquids through these walls, and flow from one parenchyma cell to the next is obviously the important concern here. Drawings of walls of parenchyma cells in wood often show perfunctory renderings in which pit borders are never indicated, and differences in pitting between end walls and lateral wall of elongate cells are not accurately illustrated. Such a basic structural component of woods as the shape and axial length of rays in wood seems not to have been subjected to interpretation. Why are rays elliptical in shape? Why do conifers lack multiseriate rays (with rare exceptions)? Axial parenchyma configurations are likewise dealt with almost exclusively at the descriptive level. The concept of “compartmentalization” by means of axial parenchyma bands (Shigo, 1984) is an attempt at interpretation, although compartmentalization related to fungus spread in wood is sufficiently infrequent (Morris & Jansen, 2016) that one may doubt that this is the sole reason for tangential banding of axial parenchyma. Shigo (1984) quite rightly feels that we should consider the role of wood structure in defense against predators. Rays and even axial parenchyma can exhibit cellular polymorphisms in cell shape, content, and abundance. We need to account for this diversification. Can we separate axial parenchyma that is involved in osmotic maintenance of water columns in woods from axial parenchyma that is involved in storage? What criteria can we use, and what methodologies are required? Separating descriptive wood anatomy and wood physiology as two different fields is very easy, but counterproductive in terms of the progress of biology. Progress in Author's personal copy

S. Carlquist biology and the other sciences has usually proceeded from fusion of fields and methodologies, not from their separation. Wood structure represents selection for optimal function, or at least selection for optimal compromises between conflicting structural requirements (Carlquist, 2017). We should be aware that all wood features have adaptive aspects, even though the goal of a particular study may not be physiological or ecological in nature. Although stability in nomenclature is to be desired, nomenclature that is meaningful in conveying biological concepts is even more desirable. This latter goal can be achieved with relatively little modification to existing nomenclature, and thereby relative stability can be achieved while synthesis of all features of woods (including physiological ones) can be accomplished. Some nomenclatural changes may be required, however.

Interpretive Data

Conifer Wood

Conifers (Fig. 1) have relatively small amounts of ray and axial parenchyma tissue compared with Gnetales (Figs. 2 and 3) and angiosperms (Figs. 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,and17). Axial parenchyma is often sparse (Fig. 1a) and rays (Fig. 1b–f) are sparse, uniseriate, and composed wholly of procumbent cells, suggesting that they are involved in radial flow of solutes. Present evidence indicates that tracheids, which average 3530 μm in length in conifers (Carlquist, 1975, Fig. 11), would be likely to contact one or more rays. The length of axial parenchyma strands in angiosperms with simple transverse perforation plates (length of axial parenchyma strands approximately equal to that of vessel elements) is about 400 μm. If axial parenchyma strands in those angiosperms are one-tenth the length of axial parenchyma strands in conifers, the number of axial parenchyma strands required to make contact to form a network in conifers is one order of magnitude smaller than that in angiosperms.

Do Any Conifers Lack Axial Parenchyma, and What Does This Indicate?

Axial parenchyma is absent in some species of Araucaria and Agathis (Greguss, 1955) and Wollemia (Heady et al., 2002) of the Araucariaceae. In Pinaceae, axial parenchyma is uncommon in Picea and Pinus (except around secretory canals) and in some species of Abies.ItispresentinCephalotaxus and Torreya but absent in the other genus of Taxaceae, Taxus. Axial parenchyma is present in Ginkgo (Greguss, 1955) and in cycads (Greguss, 1968). In all genera not named above, axial parenchyma is figured in drawings by Greguss (1955). What accounts for this situation? Axial wood parenchyma cells are not in contact with axial phloem parenchyma once a layer of tracheids has formed to separate them from the cambium, whereas wood rays are intercontinuous with phloem rays via living cells of the cambium. Therefore, the transfer of photosynthates into xylem rays is to be expected, and from rays, flow of solutes into axial parenchyma can occur (Höll, 1975). If there is less functional value for axial parenchyma than for rays in conifers, various degrees of diminution leading to absence on axial parenchyma is to be expected. The genetic information to form axial parenchyma is present in such instances, but expressed only in limited situations. Thin- Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 1 Wood features of conifers. a Chamaecyparis lawsoniana (A. Murr.) Parl., Ripon W7w, transverse section. Axial parenchyma is sparse. b–d Sequoia sempervirens Endl., Ripon W57w. b Tangential section. Tangential section, axial parenchyma contains resins. c Transverse section with latewood below, earlywood above. d Pinus strobus L., Ripon W18w. Radial section near an axial resin canal, showing thin-walled ray parenchyma. f Pinus ponderosa Douglas ex Loudon, Ripon W48w, radial section. Ray tracheids occupy nearly all of the ray shown; one file of ray parenchyma is included. Abbreviations: ap = axial parenchyma; rp = ray parenchyma file; rtp = ray-to-tracheid pitting (“cross-field pitting”); tw = thin-walled parenchyma walled parenchyma may be seen near axial secretory canals (Fig. 1e). In Pinus strobus (Fig. 1e), thin-walled ray cells are in contact with the thin-walled axial parenchyma Author's personal copy

S. Carlquist

Fig. 2 Wood features of Ephedra (Ephedraceae) and Gnetum (Gnetaceae). a–d Ephedra pedunculata Engelm. ex S. Watson, Carlquist 15,819 (RSA). a Transverse section; living fiber-tracheids (cells with contents) adjacent to vessels. b Tangential section. Rays are multiseriate and uniseriate. c Radial section. All pits on ray cell walls are bordered. d Tangential section. Living fiber-tracheids contain nuclei and vestigially bordered pits. e-f Gnetum gnemon L., Carlquist 8088 (RSA). Tangential sections of root wood. e Axial parenchyma cells containing starch. f Portions of two rays; ray cells have prominent pits; starch is evident in some cells. surrounding secretory canals. There may be a few axial parenchyma cells with lignified walls (Fig. 1e). Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 3 Gnetum gnemon, Carlquist 8088 (RSA), Wood sections. a-b Sections of root wood. a Transverse section. Parenchyma cells are filled with starch and are about equal in number to the tracheids. b Tangential section. Wide and narrow rays about equally numerous. c-d Wood from outer part of older stem. c Transverse section. Most axial parenchyma cells are located adjacent to vessels. d Tangential section. Wide and narrow rays are about equally numerous. e-f Wood of young stem. e Transverse section. Axial parenchyma cells tend to form tangential bands between the vessels. f Tangential section; Wide multiseriate rays are very tall, and narrow multiseriate rays and uniseriate rays are abundant Author's personal copy

S. Carlquist

Fig. 4 Details of ray cell pitting in angiosperms (a, b, c, e) and Gnetales (d). a Zygogynum baillonii Tiegh., Carlquist 15,609 (RSA). Radial section, showing thick tangential ray cell walls with pits mostly slightly bordered. b Laurelia sempervirens Tul., Carlquist 7223 (RSA). Radial section; pits few on horizontal and walls, but abundant and bordered on tangential and radial vertical walls. c Forchhammeria pallida Liebm., Olson 899 (UNAM). SEM of outer surface of tangential wall of a ray cell, showing numerous bordered pits. d Gnetum gnemon, Carlquist 8088 (RSA). Tangential section; starch grains visible in some cells; in others, the tangential walls with numerous pits can be seen. e Chorisia speciosa, cultivated in Santa Barbara, CA. Tangential section: pitting denser in narrower ray cells, below; pits sparser and with smaller pits or pit fields in larger cells which contain starch grains, above. Abbreviations: s = septum in septate fiber-tracheid Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 5 Pitting in axial parenchyma of angiosperms (a–e) and Gnetales (f). a Akania bidwillii (Hogg) Mabb., NSW Forestry Commission SFCw-D10096. Radial section, pits dense and mostly bordered on transverse wall of axial parenchyma strand. b Bretschneidera sinensis Hemsley, MADw-21,841. Radial section: five axial parenchyma strands and one septate fiber; pits on cross-walls of axial parenchyma are numerous and bordered. c-e Solmsia calophylla Baill., McPherson 5511 (MO). c, d SEM images of cross-walls of axial parenchyma strands in radial section, to show a single bordered pit in each. e Transverse section, outer face of cross-wall of axial parenchyma; all pits are bordered. f Gnetum gnemon, Carlquist 8088 (RSA). Tangential section, illustrating large pits on transverse wall of axial parenchyma strand. Abbreviations: ap = axial parenchyma; sf = septate fiber; tw = transverse wall Author's personal copy

S. Carlquist

Fig. 6 Diffuse and diffuse-in-aggregates axial parenchyma from transverse sections of Winteraceae (a), Canellaceae (b-e) and Sapotaceae (f). a Zygogynum baillonii, Carlquist 15,609 (RSA). Diffuse and diffuse- in-aggregates axial parenchyma. b Warburgia ugandensis Sprague (Forestry Commission of New South Wales), axial parenchyma diffuse and very scanty pararacheal (arrows point to axial parenchyma cells). c. Canella winterana (L.) Gaertn., USw-6082. Axial parenchyma abaxial. d Pleodendron macranthum Tiegh., MADw-36,444, axial parenchyma abaxial with a few laeral extensions. e Cinnamosma fragrans Baill., USw- 5502, axial parenchyma abaxial and in paratracheal bands. f Dipholis salicifolia A. DC., USw-5740. Axial parenchyma predominantly diffuse-in-aggregates Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 7 Angiosperm woods with less water storage in secondary xylem. a-b Prunus lyonii, adventive, Claremont CA. a Transverse section. Axial parenchyma is predominantly pararacheal banded. b Tangential section; uniseriate rays as abundant as multiseriate rays. c-d Fouquieria splendens. Cultivated specimen donated by Rancho Santa Ana Botanic Garden. c Transverse section, axial parenchyma in uniseriate diagonal bands, abundant in earlywood. d Tangential section; rays mostly multiseriate, but with no obvious water storage cells. e-f Hedera helix, cultivated in non-watered area, Claremont, CA. e Tangential section: rays mostly wide multiseriate, but without obvious storage cells. f Radial section: the fibriform cells are all septate fibers Author's personal copy

S. Carlquist

Fig. 8 Angiosperm woods with perceptible water storage in axial parenchyma. a-d Sections of Malva assurgentiflora (plant donated by Rancho Santa Ana Botanic Garden) that show various degrees of water storage in parenchyma. a Transverse section; axial parenchyma bands mostly narrow, paratracheal, with no obvious water storage. b Tangential section; rays large, multiseriate only, with cells large enough to permit appreciable water storage. c Transverse section from outer wood; pockets of thin-walled axial parenchyma water storage tissue present; some crystals in the thin-walled ray cells. d Tangential section with some thinner- walled axial parenchyma (left), some thicker-walled. e-f Erythrina caffra, adventive in Claremont, CA. e Transverse section. Paratracheal axial parenchyma bands composed of wide cells that account for most water storage. f Tangential section: rays narrow, small in volume compared to axial parenchyma Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 9 Angiosperm woods with minimal fiber presence. a Crepidiastrum lanceolatum (Asteraceae), Carlquist 15,679 (RSA). Transverse section; libriform fibers are relatively scarce compared to axial parenchyma. b–d Chorisia speciosa (bombacoid Malvales), cultivated, Claremont, CA. b Ray cells from tangential section; larger cells contain starch, narrower cells have dense pitting on end walls. c Transverse section. Libriform fibers are much less common than the relatively large axial parenchyma cells. d Tangential section. Margins of rays merge imperceptibly into axial parenchyma; packets of narrow radially elongate cells occur in the rays. e-f Plumiera alba (Apocynaceae), cultivated in Claremont, CA. e Transverse section. Axial parenchyma cells large, fibers isolated or in tangential groups. f Tangential section; rays narrow, composed mostly of procumbent cells distinguishable from the axial parenchyma Author's personal copy

S. Carlquist

Fig. 10 Angiosperms with water-storage wood and raylessness. a–b Jacaratia hassleriana Chodat (Caricaceae), Arid Lands Greenhouses. a Transverse section; secondary xylem of stem, fibers lacking in xylem; fibers present in secondary phloem at top. b Tangential section of stem; secondary xylem consists wholly of water storage tissue except for the vessels. c-d Crassula argentea, cult. Claremont, CA. c Transverse section; axial parenchyma in fascicular secondary xylem, wide ray at right contains tannin idioblasts. d Tangential section, ray portion at right; vessels have helical thickenings. e-f Kalanchoe beharensis Drake (Crassulaceae), cultivated in Santa Barbara, CA. e Transverse section of wood; vessels occur in small clusters that contain axial parenchyma. f Tangential section; wood is rayless Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Does Axial Parenchyma in Conifers Function in Supporting Conduction as it Does in Angiosperm Woods?

The distribution of axial parenchyma in angiosperms supports the idea that conduction in vessels is supported or maintained by the action of adjacent axial parenchyma cells (see references cited in Introduction). This does not seem likely in conifers because of sparseness or absence of axial parenchyma in an appreciable number of conifers. In fact, one can ask what, if the all-tracheid secondary xylem of conifers is so successful that hydraulic safety margins in conifers are so much greater in conifers as a whole than in angiosperms (Hacke et al., 2015), is the function of axial parenchyma in conifers that have it? The quantities seem insufficient for storage or photosynthates, so that one is inclined to look for subsidiary functions. Axial parenchyma cells in conifers often fill with resins or other secondary compounds rapidly, suggesting a rot or predator deterrence. The potential function of axial parenchyma in regeneration following stem injury is difficult to demonstrate. Axial parenchyma in conifers can be confined to the terminal portions of growth rings, but more commonly is scattered in a diffuse pattern throughout the ring (Fig. 1a). Axial parenchyma in conifers can be rapidly converted to resin-containing cells (Fig. 1c) in growth rings that are no longer actively conducting, another evidence of the limited function of axial parenchyma in conifers. An axial parenchyma that is subdivided, but with bordered pits, shown in Fig. 1d, exemplifies the fact that axial parenchyma is more diversified in conifers than most studies would lead us to believe.

What Do Rays in Conifers Do?

Rays in conifers (other than Gnetales) are rarely more than one cell in thickness (Fig. 1b; Greguss, 1955). This does not seem compatible with a storage function, and mentions of starch presence in rays are virtually non-existent. Resins or similar compounds are often seen in conifer ray cells, however (Fig. 1b), just as they are, simultaneously, in axial parenchyma (Fig. 1a–d). Some ray cells can remain devoid of contents. The tangential walls of such cells are often thin, or bear large pits, simple or bordered, (“nodular,” or “nodulated”) with intervening secondary wall portions that define the large size of the pits. There are accurate drawings of the circular and oval pits on the tangential walls of ray cells of Cupressaceae, etc. in Greguss (1955) and similar SEM images in Roman-Jordan et al. (Roman-Jordan et al., 2016,Figs.6c and 9c). These cells could certainly serve for flow of solutes. The presence of such large pits between ray cells and tracheids would be difficult to explain otherwise. That rays in conifers have cell polymorphism and therefore more than one function is clearly demonstrated by the occurrence of ray tracheids in Pinus and Abies (Holden, 1913;Greguss,1955) as well as a few Cupressaceae (Roman-Jordan et al., 2016). Ray tracheids have been reported in one angiosperm, Tetracentron (Kedrov, 2012). The conifers without ray tracheids seem more characteristic of tropical or moist habitats as a whole, but there are exceptions. Ray tracheids as a pathway for sap flow suggests a role in transferring water from one tracheid to another. One might be able to answer this question with the use of dyes. Conifer rays can contain secretory canals (Picea, Pinus, Pseudotsuga). This seems clearly a use of rays as a defense mechanism for the wood, similar to the axial secretory Author's personal copy

S. Carlquist

canals in wood and the abundant amount of resin secretion to be seen in conifer bark. Defense mechanisms, such as crystals, tannins, etc., are often more abundant closer to the surface of a stem (or other organ) than further inside. However, we should note that the expenditure of space and photosynthates on resin canals and on resin-containing ray cells and axial parenchyma is larger than on cells that could be claimed to be involved Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

R Fig. 11 Woods of Cactaceae (a-e); wood with successive cambia (f-g). a-b Pereskia aculeata, cult. In Claremont, CA. a Transverse section, starch present in axial and ray parenchyma. b Tangential section; axial parenchyma cells (non-subdivided), caner’ libriform fibers, left. c-e Cereus repandus, cultivated in Claremont, CA. c Transverse section. Two files of secondary xylem fibers have changed to production of axial parenchyma. d Tangential section; portions of two large multiseriate rays. e Tangential section, higher power; limits between ray and axial parenchyma are uncertain. f-g Stegnosperma halimifolium Benth., Carlquist s. n., 1969. f Transverse section of outer portion of stem, to show two vascular increments and the bands of parenchymatous conjunctive tissue that alternate with them. g Ontogenetic origin of successive cambia and conjunctive tissue from a master cambium. Abbreviations: ap = axial parenchyma; c = cortex; ct = conjunctive tissue; mc = master cambium; mc (d) = location of the dormant master cambium; r = ray; sc = secondary cortex; sp. = secondary phloem; sx = secondary xylem

in flow or storage of some sort. Conifers are capable of producing wider rays and more axial parenchyma where resin canals occur. Parenchyma cells in conifer woods may play less of a role in the osmotic regulation of conduction than they do in angiosperms. However, various types of “cross-field pitting” (pits between tracheids and ray cells, Fig. 1e) lead us to believe that ray cells do have large pit areas in contact with tracheids, and these contacts are likely to be involved in some kind of sugar and/or ion interaction in growth rings that are actively conducting. Storage of starch in conifer rays is scarcely mentioned in literature on conifer woods. These facts tell us much about the efficiency of the conifer tracheid, suggesting than it requires less of a conduction support system that do angiosperm vessels, which are relatively wide. The diameter of conifer tracheids must be relatively small in order to achieve resistance to embolisms and recovery from freezing (Davis et al., 1999). More numerous tracheids of relatively small diameter (compared to vessels of angiosperms) form a basic principle of conifer woods. There is a potential gain in various types of wood strength if rays are relatively small. Strength in mechanical cells of wood is conferred by helical cellulose microfi- brils in a lignified background in secondary walls, but also, quite significantly, by cementing substances between fibriform cells. The relatively great length of fibriform cells increases their adhesive surface (Wellwood, 1962). Thus, the interruption of fibriform tissue by rays would decrease strength. Indeed, we can see this macroscop- ically in samples of dried angiosperm woods, where the surfaces of shrinkage cracks run parallel to rays (“checking”). Thus, there is a negative selective value for large (especially tall) rays in woods. Such rays are most prominent in the stems of lianas, where flexibility is at a premium compared to trunks of larger trees. In this perspective, space in conifer woods is better devoted to more tracheary elements than to larger rays and more axial parenchyma, which thereby play a modally different role in conifers than they do in angiosperms.

Gnetales

The inclusion of data from Gnetales is important to our understanding of hydraulics in conifers and angiosperms. Long ago, Thompson (1918) presented reasons why Gnetales should not be placed close to angiosperms, a conclusion accepted by Bailey (1944) and confirmed on the basis of molecular data by Bowe et al. (2000), Burleigh and Mathews (2004), and subsequent authors. Thus, conifers plus Gnetales become a group that has been called “gnetifers” or “gnepines.” Wood of Gnetales contains Author's personal copy

S. Carlquist

virtually all of the characters found in angiosperm woods (Carlquist, 2012b), in contrast with conifer wood. Thus, wood of Gnetales offers us an excellent answer to the question: what would happen to parenchyma if vessels were evolved in a vesselless gymnosperm? Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

R Fig. 12 Angiosperm woods showing cell type dimorphism. a-e Ficus elastica (Moraceae), cultivated in Claremont, CA. a Transverse section; fiber bands alternate with wide axial parenchyma bands which are neither apotracheal nor paratracheal. b Tangential section; axial parenchyma at left, libriform fibers at right. c Radial section; axial parenchyma strands at right, vessel-adjacent parenchyma cells at right are small and contain gummy (latex?) droplets. d Enlarged transverse section to show starch-rich cells of axial parenchyma band; recent divisions evident in vessel-adjacent parenchyma. e Radial section; point of intersection between axial parenchyma and ray. f Eryngium bupleuroides Hook. & Arn. (Apiaceae), USw-33,857, Transverse section; earlywood fibers are living, latewood fibers are non-living. g Robinia pseudoacacia L., (Fabaceae) Ripon W380w, transverse section. Fibers that are sectioned at their widest point appear like axial parenchyma, but are not

Ray and Axial Parenchyma of Gnetales: Details Related to Function

The parenchyma features by which Gnetales differ from conifers include presence of both multiseriate and uniseriate rays (Figs. 2b and 3b, d, f). Borders are present on pits of ray cell walls (Fig. 2c, f)ofEphedra and Gnetum.Bordered pits are also present on axial parenchyma of Gnetum (Fig. 5f), although simple pits with wide pit membrane areas predominate. Small borders occur on pits of the living fiber-tracheids of Ephedra, which are functionally equivalent to the axial parenchyma of Gnetum (Fig. 2d). Similar pits occur on septate fibers of Gnetum (Fig. 2f and 4d). Some, or a large proportion of the axial parenchyma cells of Gnetum (Fig. 3a,c,e)aswellasthelivingfiber-tracheidsofEphedra (Fig. 2a) are associated with vessels. Septa sometimes occur within the axial parenchyma strand cells of Gnetum (new report). Welwitschia has thin-walled axial parenchyma and thin-walled ray cells, both with simple pits (Carlquist and Gowans 1995). In Gnetum gnemon, rays in wood of roots (Fig. 3b) are similar to those in the adult stem (Fig. 3d), whereas those in wood of a young stem (Fig. 3f) are taller, with uniseriates more common. These features of the young stem wood rays accords with well-known earlier stages in ray ontogeny (see Barghoorn, 1941; Carlquist, 1988)in angiosperms, as illustrated in those two sources for Bursera simaruba. We can even refer the rays of Gnetum gnemon to the Heterogeneous Type I of Kribs (1935), whereas the rays of Welwitschia secondary xylem (Carlquist and Gowans 1995)correspondto Paedomorphic Type I of Carlquist (1988). The juvenilistic rays of Welwitschia are related to the fact they are produced within successive cambia, each of which has limited duration, although the vascular cambium of each of the numerous vascular increments does produce secondary phloem and secondary xylem.

Parenchyma in Wood of Gnetales: Indicators of Photosynthate Flow or Storage

Starch in rays (Figs. 2c, f and 4d), and in septate fiber-tracheids (Fig. 2e)isseeninthe Gnetales, as it is in angiosperms. In Gnetales, the nature of pitting, ray cell size, and starch presence does not suggest a strong differentiation into flow cells versus storage cells. Perhaps the ray cells of Ephedra (and living fiber-tracheids in Ephedra) can serve both purposes, changing on a seasonal basis. The storage of starch in roots of Gnetum gnemon is considerable (Figs. d2e an 3a): axial parenchyma cells are more common than tracheids. No conifer has been reported to have such a high proportion of root wood cells devoted to storage. Author's personal copy

S. Carlquist

The abundant starch storage in rays, axial parenchyma, and septate fiber- tracheids of Gnetales is clearly different from the minimal presence of starch storage in conifers, but Gnetales are quitesimilartoangiospermsinthestorage of starch in living cells. This suggests that Gnetales can store starch in relation to Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

R Fig. 13 Transverse sections of stem wood of Brassicales to show unusual axial parenchyma distributions. a-b Reseda alba L., Kennedy March 2, 1973 (POM). a Axial parenchyma associated with earlywood. b On the same section, axial parenchyma associated with latewood. c Caylusea hexagyna M. L. Green, Podlech 42,683 (RSA). Axial parenchyma is both apotracheal and paratracheal. d Stanleya pinnata Britton, cult. Rancho Santa Ana Botanic Garden, axial parenchyma band in both latewood and earlywood. e Tersonia brevipes Moq., Carlquist 5385 (RSA), axial parenchyma abundant in earlywood; elsewhere diffuse or diagonal diffuse-in- aggregates. f Salvadora persica L., cult. Univ. Calif. Santa Barbara greenhouses. Arrows indicate vascular cambium; Interxylary phloem strand located in axial parenchyma band. Abbreviations: ew = earlywood; ixp = interxylary phloem; lw = latewood; sp = secondary phloem

flushes of growth and fruiting. More importantly, perhaps, living cells accompany vessels and tracheids in Gnetales, so there is a potential source of carbohydrates that could serve for transfer into the water columns of vessels. We can see this association in Ephedra.InFig.2a, living fiber-tracheids (identifiable by cytoplas- mic contents) can be seen above the larger vessel. The fiber-tracheids are more common around vessels than in the tracheid background. In Gnetum gnemon,axial parenchyma is also mostly vasicentric in older stems (Fig. 3c), although some diffusely distributed axial parenchyma stands can easily be found. In the young stems of Gnetum gnemon (Fig. 3e), axial parenchyma, identifiable by thinner walls, tends to form tangential bridges between vessels, not unlike paratracheal banded axial parenchyma of angiosperms. Studies on wood physiology of Gnetales are effectively non-existent. The reasons for this presumably lie in the lack of commercial value of Gnetales. Gnetales are not difficult to grow, so such studies should be attempted, because the implications will probably go well beyond the Gnetales, to conifers and to angiosperms. Likewise, we need studies on the physiology of wood of cycads, which have multiseriate rays and axial parenchyma (Greguss, 1968), and also on wood of Ginkgo, which has both uniseriate rays and axial parenchyma (Greguss, 1955).

Angiosperms

Functions of the Parenchyma Network Listed

If conifer phylogeny has involved minimization of the parenchyma network in woods, whereas angiosperms feature increased quantities of parenchyma, there is the possibility of multiplication of functions in angiosperm wood parenchyma. In order to look for anatomical signals of these functions, we first should have an idea of what functions might be served. The following list represents the underpinning of anatomical information reported in the balance of this paper.

(1) Conduction of photosynthates in solution. (2) Storage of photosynthates, chiefly as starch. (3) Control of conduction by transfer of sugars and ions into the apoplastic stream. (4) Water storage. (5) Structural support by means of turgor. (6) A subsidiary source of mechanical strength by means of secondary walls; thinner ray cell walls aid in expansion and contraction of stems. (7) Closure of vessels (especially in earlywood) by means of tylosis formation. Author's personal copy

S. Carlquist

(8) Sites for adjustment to torsion (e.g., rays in liana stems). (9) Auxiliary cells in latewood aiding survival of cambium. (10) Sites for beginning of tissue regeneration following injury. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

R Fig. 14 Septate and living fibers in angiosperm woods. a-b Trattinckia demerarae Sandwith (Burseraceae), MADw-19,938. a Transverse section; background cell type is septate fibers; axial parenchyma identifiable as slightly wider cells around the vessels. b Radial section. Axial parenchyma strands (lower left) can be distinguished from ray cells (top) and septate fibers (lower right and center). c Nerium oleander (Apocynaceae), cult. Claremont, CA. Tangential section. All fibers are septate. d Foeniculum vulgare Mill. (Apiaceae), adventive in Santa Barbara, CA. Fibers are living (some septate), rays tall, juvenilistic. e-g Pimpinella dendrotragium Webb & Berthel. (Apiaceae), Carlquist 2709 (RSA). e Transverse section. Rays absent. f Transverse section, higher power: axial parenchyma cells surround vessels (some recently divided, walls thinner than those of fiber). g Tangential section; section entirely rayless except for the two small uniseriate rays shown

Ray Parenchyma Cell Walls Reveal Modes of Flow in Living Cells

Ray cells commonly are much more prominently pitted on tangentially-oriented walls than on walls that have radial (either vertical or horizontal) orientation. This is shown by all of the photographs of Fig. 4. The pits, when secondary walls are thick enough to aid examinations, and when seen in sectional view, may be simple or bordered (a dumbbell shape of wall segments between pits in sectional view is a key to presence of borders). Particular attention shouldbepaidtoFig.4b, in which one can compare the prominently pitted tangential walls (vertically oriented in the photograph, seen in sectional view) with the sparsely pitted horizontal radial walls (oriented horizontally in the photograph) and the sparsely pitted vertical radial walls (in face view in the photograph). The pits in Fig. 4b are all bordered. One may also see borders on ray cell walls by SEM imaging of the outer surfaces of ray cells (Fig. 4c). The presence of borders, as in borders on cells of the apoplastic water-conduction system (vessels, tracheids), shows a combination of wide pit membranes with overarching pit borders. The wide pit membranes maximize conduction across the pit membrane, the borders minimize loss of mechanical strength (Carlquist, 1988, 2017), a compromise between two opposed requirements. Bordered pits on ray cells have been held to indicate flow (Carlquist, 2007a). The appreciation of bordered pits on ray cells has lagged because wood students are accustomed to viewing bordered pits in face view, as in tracheids and vessels, rather than in sectional view. The thick tangential walls of ray cells make detection of borders in face views of pits difficult, although they are clearly revealed with SEM preparations that expose the outer tangential surface of a ray cell (Fig. 4c). Sizes of pits in ray cells can also be assumed to indicate adaptation to flow. Ray cells of Gnetum (Fig. 4d) are used to illustrate this for convenience, but the same phenomenon can readily be seen in rays of angiosperm woods. Pits on tangential walls of ray cells, especially radially elongate ray cells, are frequently much larger than those on the radial walls, indicative of probable direction of flow. Density of pitting on ray cell walls can also be assumed to be indicative of adaptation to flow. Chorisia speciosa (Fig. 4e) has pits densest on the narrow tangential walls of radially elongate ray cells, grading to medium density in medium-diameter cells (Fig. 4e, lower right), and to sparse in larger cells. The pitting in walls in the larger ray cells in Fig. 4e are not shown, but the presence of starch grains is.

Axial Parenchyma Cell Walls Often Reflect a Conductive Function

Certainly pits are more prominent on ray cells than on axial parenchyma cells. If one looks at the transverse walls between the stand cells of an axial parenchyma strand, one Author's personal copy

S. Carlquist

can see that these walls bear pitting more pronounced than that of the axial (vertically- oriented) walls of axial parenchyma. This can be seen in radial sections of wood of two genera of Akaniaceae, Akania (Fig. 5a)andBretschneidera (Fig. 5b). In Akania,the pits on the end walls (sectional view, upper right) are much more numerous than those of the vertical walls (below the cross wall, in face view). All of these pits have borders, Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

R Fig. 15 Wood functions related to axial parenchyma in Apiales. a-f Foeniculum vulgare, adventive in Santa Barbara, CA 93110. a Primary xylem pole of a main vascular bundle in transverse section; protoxylem elements collapse, leaving uniseriate rings of axial parenchyma. b Area between main bundles in which secondary xylem forms from interfascicular cambium without primary xylem formation. c Transverse section of secondary xylem; a narrow band of latewood is present. d Radial section of terminal latewood. e A vascular tracheid from the terminal latewood, from a radial section. f Metaxylem vessel with an associated strand of axial parenchyma (left). g Oreopanax steinbachianus Harms, cult. Santa Barbara, CA. SEM image of radial section of stem, primary xylem at left, secondary xylem at right. Abbreviations: ew = earlywood; lw = latewood; ptx = protoxylem; px = primary xylem; sf = starch in septate fiber; sx = secondary xylem

albeit inconspicuous ones. In the Bretschneidera section (Fig. 5b), transverse walls of five parenchyma strands are shown. In each of these diagonal cross-walls, bordered pits are densely placed. Pits are much sparser (but slightly bordered) on lateral walls of these strand calls. Solmsia of the Thymeleaceae was selected for SEM study of axial parenchyma pitting because it has thick axial parenchyma walls. Bordered pits in sectional view occur on the transverse walls of axial parenchyma strands (Fig. 5c, d). As seen in face view of the outer wall surface (from a transverse section of the wood). The bordered nature of the pits is clear (Fig. 5e). The transverse wall of an axial parenchyma strand in wood of Gnetum gnemon is shown in Fig. 5f. Pit membrane areas are more extensive in the transverse wall than are the secondary wall portions. This condition can often be seen in angiosperms. Some- times the transverse wall of an axial parenchyma strand bears a single large pit.

Axial Parenchyma: Changes in Patterns of Aggregation and Abundance

Although Amborella is considered the sister to the remainder of the angiosperms in all recent molecular phylogenies, particular features of this monotypic genus may or may not be plesiomorphic. However, Amborella has sparse diffuse axial parenchyma (Carlquist, 2012a), as do a number of other angiosperms with a large number of features now interpreted a plesiomorphic (Metcalfe & Chalk, 1950; Metcalfe & Chalk, 1983; Metcalfe, 1987). Although one can say that the axial parenchyma of Amborella, or for that matter, Warburgia (Fig. 6b) is too scarce to function in the parenchyma network, one must remember that three-dimensional wood studies (Zimmermann, 1971; Kedrov, 2012) indicate that living cells in wood are not isolated. Also, the tracheids in these genera are long, so more numerous axial-parenchyma-to- tracheid contacts are likely to occur than study of transverse sections might suggest. One notices in woods that have “diffuse” axial parenchyma that some axial parenchyma cells do contact vessels (Fig. 6b) and some form tangential groupings (Fig. 6a, f). Truly random distributions of axial parenchyma are rare—andeveniftheyoccur, some of the axial parenchyma strands would contact vessels here and there. The tangential groupings, commonly called diffuse-in-aggregates (Fig. 6a, f), rarely form long continuous tangential bands, nor do bands two or three cells thick. Thereby, tangential lines or bands of axial parenchyma do not seem likely to compartmentalize pockets of fungus invasion, as envisioned by Shigo (1984), although that may be operative in certain woods. One can ask why, when vessels are mostly radially grouped in angiosperm woods, banded axial parenchyma runs not parallel to those bands but at right angles to them. The most obvious explanation would be that upright cells Author's personal copy

S. Carlquist

Fig. 16 SEM images of tangential (a-d) and radial (e-f) sections of young stem of Oreopanax steinbachianus (cult. Santa Barbara, CA). a Outer (left) and inner (right) surfaces of ray cells. Pits are bordered. Inner surface of ray cell; pits of various sizes, some subdivided. c Starch storage cell (left) and flow cell (right) from tangential section of ray, showing contrasting sizes and densities of pits. d Tip (wing) cell of ray from tangential section of ray; as seen from outer surface (below), pits are minute and bordered) inner surface of cell with starch, above. e Radial section, procumbent (flow) cells above, wing cell file below; pits on outer surfaces of cells are minute, sparse. f Metaxylem—secondary xylem transition in radial section as indicated by progressive dwindling of bar number in perforation plates. Axial parenchyma surrounds all vessels belonging to wings of multiseriate rays or in uniseriate rays do run parallel to radial chains of vessels. Radial chains of axial parenchyma would therefore be redundant. Winteraceae (Fig. 6a) and Canellaceae (Fig. 6b–e) are sister families according to recent molecular phylogenies, so they give us a chance to compare woods with vessels (Canellaceae) with vesselless woods (Winteraceae). One unexpected result is that the Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... vesselless wood (Fig. 6a) has tracheids much wider than the tracheids or fiber-tracheids in the vessel-bearing woods (Fig. 6b–e). Thus, the conductivity of the tracheids in most Winteraceae is much greater than one would have supposed (all of the photographs on Fig. 6 are at the same scale of magnification). With the potential support of the tracheids by diffuse and diffuse-in-aggregates axial parenchyma, a vesselless wood is thereby not disadvantaged. Although Metcalfe (1987) figures more abundant diffuse axial parenchyma (and definitely tracheids) for Warburgia stuhlmannii Engl., some axial parenchyma cells contact vessels in his figures as well as mine (Fig. 6b). In Canella (Fig. 6c)and Pleodendron (Fig. 6d), the axial parenchyma is clearly abaxial (a form of paratracheal parenchyma), with some lateral extensions in Pleodendron.InCinnamosma (Fig. 6e), the axial parenchyma is transitional between paratracheal and paratracheal-banded. Thus, the wood anatomy of Canellaceae features repatterning of the axial parenchyma so that the arrangement of axial parenchyma strongly suggests a functional interaction with vessels. The most probable one at this moment seems to be the transfer of sugars into vessels as a way of maintaining the safety of water columns. Although Canellaceae have tracheids (possibly sometimes fiber-tracheids), as clearly shown in the Metcalfe (1987) figure for Warburgia stuhlmannii, Dipholis of the Sapotaceae (Fig. 6f) has fiber-tracheids. Thus, although there is an association between tracheids and diffuse parenchyma in the majority of woods with tracheids, there are exceptions. One should note that the lines of diffuse-in-aggregates axial parenchyma and diffuse cells in Dipholis contact both vessels and rays in places. Thus, just an occasional axial parenchyma cell or two in contact with a vessel may serve for maintenance of water columns. In fact, axial parenchyma in Asteraceae is mostly scanty paratracheal (scanty vasicentric), also illustrating that whatever the function of axial parenchyma may be in relation to vessels, the quantity of axial parenchyma cells touching a vessel or vessel group can be rather minimal. Diffuse parenchyma should not be viewed as an inefficient way of contact between the parenchyma network and vessels. Rather, a relatively high degree of association between diffuse axial parenchyma and presence of tracheids (Kribs, 1937)suggeststhat diffuse axial parenchyma is supporting conduction processes in both tracheids and vessels of a given wood. Diffuse-in-aggregates very likely has the effect of increasing the number of contacts between rays and axial parenchyma as well as among axial parenchyma cells.

The Anatomy of Water Content in Woods

Comparisons of water content in freshly harvested (“green”) conifer and angiosperm woods are few. One is the USDA (1974) compilation, based largely on results from the Forest Products Laboratory (J. F. Siau, 1995, personal communication). According to Table 3–3 in that reference, sapwood water percentages are higher in conifers than in angiosperms, probably because at the time of harvesting, tracheids in conifer woods retain water, whereas in angiosperm woods, vessels empty and fibrous tissue (libriform fibers mostly) is less likely to contain water. In the list of the woods in Table 3–3 (USDA, 1974), none of the woods would be considered succulent. That would not be expected in a publication concerned with commercially usable woods. Within the angiosperms listed in that table, higher percentages of water content were reported in Author's personal copy

S. Carlquist

angiosperms of wetter habitats: Betula, Liquidambar, Magnolia, Nyssa, Platanus,and Populus. Half of those six genera have tracheids as their imperforate tracheary elements, so water may be retained in those tracheids upon harvesting. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Fig. 17 Rays of Apiales (a-e), Chloranthaceae (f-g) and Campanulaceae (h). a-g. Radial sections to illustrate ray cell shape. a. Myodocarpus fraxinifolius Brongn. & Gris (Myodocarpaceae), Carlquist 4268 (RSA). Ray cells are radially elongate, typical of the basal woody Apiales (e.g., Araliaceae). b Heteromorpha trifoliata Eckl. & Zeyh. Apiaceae), Carlquist 2670 (RSA). Ray cells are mostly square to upright. c. Eryngium inaccessum Skottsb., (Apiaceae), Skottsberg 20 (Göteborg Botanic Garden). A few files of procumbent cells in center of ray, ray otherwise of square and upright cells. d-e Arracacia atropurpurea Benth. & Hook. f. (Apiaceae), Iltis 277 (MAD). d Secondary xylem close to pith; ray cells are all upright. e Secondary xylem close to cambium; several files of procumbent cells, other ray cells are square to upright. f Hedyosmum scabrum Solms, SJRw-2864, radial section. All ray cells are upright. g Chloranthus officinalis Blume, Stone 12,116 (KL), radial section. All ray cells are upright. h Cyanea leptostegia A. Gray, Carlquist 1961 (RSA), tangential section. Rays are very tall, composed wholly of upright cells

Water content in typically woody angiosperms is relatively small compared to that in angiosperm wood and bark which could be considered to have degrees of succulence. In Table 1, the values reported for “WM/WD” for typically woody angiosperms are very similar to those reported in Table 3–3ofUSDA(1974). However, the species in Table 1 were selected as a ways of comparing angiosperm species with wood probably devoted to sequestering relatively large quantities of water to the typically woody species. To develop the data of Table 1, stems of the selected species were harvested between 9 and 10 AM between January 10 and 13, 1986. The plants had received only rainwater since November 1 of 1985, but that season was above average in rainfall in Claremont, California. Bark was separated from secondary xylem and pith, and bark and wood were weighed separately fresh and after drying in a 60 °C oven. Values are presented as percentages in Table 1 in order to reveal the relative roles of bark and wood in water storage in the various species. Microtome sections were also prepared of the woods listed in Table 1 as a way of determining the probable sites of water storage within the woods. From those, examples were selected for illustrations (Figs. 7, 8, 9, 10, 11, 12, 13,and14). We find that sites for water storage are quite diverse in the more succulent (boldface) species studied. Not included in this study is the effect of seasonal stem expansion and contraction that is covered by Scholz et al. (2008). That study included bark as well as wood.

Characters of Non-succulent Stems

In Table 1, non-succulent bark is indicated for Artemisia californica, Bougainvillea glabra, Cercidium floridum, Erythrina caffra, Fouquieria splendens, Hedera helix, Malosma laurina, Malva assurgentiflora, Nerium laurina, Peritoma (Isomeris) arborea,andPrunus lyonii. Fouquieria splendens, a desert shrub, is often thought of in conjunction with succulents, but its bark features thick sclerenchyma layers. In fact, most desert shrubs other than cacti do not have much water storage in bark. Instead, drought deciduousness is the most common way of dealing with the dry season.

Succulent Bark

Bark relatively high in water storage is shown in Aeonium arboreum, Cereus repandus, Chorisia speciosa, Crassula argentea, Euphorbia pentagona, Ficus elastica, Nicotiana glauca, Leptosyne (Coreopsis) gigantea, Pereskia aculeata, Portulacaria afra,and Ricinus communis. About half would traditionally be classified as succulents (the species of Aeonium, Euphorbia, Leptosyne, Pereskia,andPortulacaria), but the Author's personal copy

S. Carlquist

Table 1 Water content of bark and wood in angiosperm species

Species BF/BD BM/BD BM/WM BF/WF WF/WD WM/WD WM/WF

Aeonium arboreum Webb & Berth. 7.32 6.32 9.38 7.06 2.85 1.85 0.65 Aesculus californica Nutt. 3.27 2.27 0.34 0.26 2.21 1.21 0.55 Artemisia californica Less. 1.90 0.90 0.37 0.20 1.59 0.59 0.37 Bougainvillea glabra Choisy 2.78 0.64 0.38 0.22 1.60 0.59 0.37 Calliandra inaequilatera Rusby 2.41 1.41 0.68 0.68 1.63 0.62 0.38 Cercidium floridum Benth. 1.83 0.84 0.19 0.16 1.60 0.60 0.37 ex A. Gray Cereus repandus Mill. 9.14 8.14 5.62 5.09 5.14 4.14 0.85 Chorisia speciosa A. St. Hil. 4.16 3.16 1.43 1.17 2.63 1.63 0.62 Crassula argentea Thunb. 8.92 7.92 4.74 4.51 11.90 10.95 0.92 Erythrina caffra Thunb. 3.66 2.66 2.34 1.43 3.93 2.93 0.74 Euphorbia pentagona Haw, 6.25 5.25 1.33 1.05 2.99 2.93 0.67 Ficus elastica Roxb. ex Hornem. 4.10 3.11 1.40 1.00 2.19 1.19 0.54 Fouquieria splendens 2.40 1.40 1.13 1.39 1.59 0.59 0.37 Engelm in Wisl. Hedera helix L. 2.78 1.79 0.23 0.19 2.10 1.10 0.52 Leptosyne gigantea Kellogg 8.92 7.92 1.93 1.50 4.55 3.49 0.78 Malosma laurina Engl. 2.43 1.43 0.45 0.31 1.66 0.66 0.40 Malva assurgentiflora 3.86 2.86 0.80 0.73 3.14 2.14 0.68 (Kellogg) M. Ray Nerium oleander L. 2.18 1.17 0.56 0.43 1.72 0.74 0.72 Nicotiana glauca Graham 5.16 4.14 0.40 0.29 2.33 1.33 0.57 Pereskia aculeata Mill. 4.44 3.44 1.14 0.80 2.23 1.23 0.55 Peritoma arborea (Nutt.) Iltis 2.02 1.01 0.21 0.24 2.27 1.28 0.56 Plumeria alba L. 6.64 5.64 1.69 1.24 2.66 1.66 0.62 Portulacaria afra Jacq. 4.69 3.69 0.67 0.56 2.96 1.98 0.66 Prunus lyonii Sudw. 2.60 1.16 0.31 0.16 1.47 0.46 0.32 Ricinus communis L. 4.68 3.66 0.61 0.50 2.80 1.80 0.65

Higher values for each column are in boldface, and indicate greater water content; lower values in each column are in italics and indicate lower water content. Columns 3 and 4 indicate degree of allocation of water between bark and wood. B = bark; D = dried; F = Fresh (weight); M = moisture (difference between fresh weight and dry weight); W = wood. Further explanations in text exceptions are of special interest. Ricinus communis would probably be considered a semi-succulent shrub or small tree, and Nicotiana glauca would probably also fall into this category. The idea that Ficus elastica has water storage in its bark should not be surprising, because Ficus is prominent in dry tropical localities. Malva assurgentiflora falls only slightly below the species in the “succulent bark” category; it grows on maritime rocky shores and is thus a marginal halophyte. The succulence of a coastal halophyte such as Cakile edentula (Bigel.) Hook. is somewhat more pronounced. The third and fourth columns of Table 1 compare the water content of bark to that of wood for the species studied. These figures highlight concordance or disparity between bark and wood for the species studied. Disparity is most evident for Aeonium Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... arboreum, which has a fiber-free succulent bark but a dense, fibrous rayless wood, much like that of Kalanchoe beharensis (Fig. 10e, f), also in Crassulaceae. The tree species of Erythrina have prominent water storage in secondary xylem, but not in bark. The species in Table 1 in which neither bark nor wood is adapted for water storage include Aesculus californica, Artemisia californica, Bougainvillea glabra, Cercidium floridum, Hedera helix, Malosma laurina,andPeritoma arborea. These species grow in dryland situations, but can be regarded as truly woody species. Aesculus californica, Artemisia californica,andCercidium floridum are all drought-deciduous.

Water Storage in Secondary Xylem

Figure 7 is devoted to species that show no evidence of enhanced water storage in wood. These are all species included in Table 1. Prunus lyonii (Fig. 7a, b) has bands of axial parenchyma, not exceptional in amount for a woody species; the bands intersect most vessels or vessel groupings as seen in transverse section. Rays in this species do not occupy a large proportion of the wood as seen in tangential section (Fig. 7b). The cells are of moderate size and thick-walled, and thus are not suited for water storage to any pronounced degree. Ray cells have rigid secondary walls (Fig. 7a, b)whichwould be antithetical to the demands of water storage, in which a cell volume fluctuates. Axial parenchyma is similar in wall thickness, but the cells are even smaller. The wood of Fouquieria splendens (Fig. 7c, d) is similar to that of Prunus lyonii. The rays (Fig. 7d) are larger than those of Prunus lyonii, but when compared to those of Malva assurgentiflora (Fig. 8b, they are relatively small, and occupy a smaller proportion of the wood. The axial parenchyma of F. splendens is more abundant in earlywood than in latewood, and most of each growth rings is latewood (Fig. 7c). Axial parenchyma in latewood occurs in groupings that look diagonal more often than tangential. In any case, the total volume of axial parenchyma and its cell size in F. splendens mitigate against this tissue serving as a locus for water storage. The short shoots of F. splendens do leaf out very rapidly (in a few days or a week), a habit that correlates with the greater volume of rays compared to those of P. lyoni. Hedera helix (Fig. 7e, f) has large rays, composed of procumbent cells (Fig. 7e). Such rays can serve as massive conduits for photosynthates in solution that are required for the flushing habit characteristic of Araliaceae (the upright adult shoots of Hedera show this much more clearly than the sprawling juvenile shoots). Axial parenchyma is not prominent. Instead, the background of Hedera wood, as in other Araliaceae, is composed of septate fibers that can be shown to store starch prior to growth events (see Fig. 15g). Thus, the wood of Hedera is not a prominent source of water storage, but is instead devoted to photosynthate storage and retrieval.

Diverse Designs for Water Storage in Wood

Malva (Lavatera) assurgentiflora wood (Fig. 8a-d) shows some wood succulence, a degree confirmed by the figures for it in Table 1. Axial parenchyma is banded, mostly paratracheal but with some apotracheal patterning as well (Fig. 8a). The bands in Fig. 8a are not unusually abundant, but some stems of this species, especially older ones, tend to show radially wider bands of axial parenchyma with thin walls (Fig. 8c, d). Thus, M. assurgentiflora has a flexible system for devoting more prominence to Author's personal copy

S. Carlquist mechanical tissue in younger upright stems while producing more abundant water storage tissue at the periphery of outer stems and at the bases of branches and in roots. Wall thickness varies in axial parenchyma, presumably in accordance with shift towards storage and away from mechanical strength, of vice versa. The contacts between axial parenchyma bands and rays are massive (Fig. 8c). The amount of ray tissue in M. assurgentiflora is much greater than in the non-succulent species or Fig. 7 (compare with Fig. 8b, d). The rays are also taller than those of a typical woody plant. The rays are composed of three kinds of cells: (1) most common, isodiametric cells with no visible contents; (2) similar cells with rhomboidal calcium oxalate crystals, presumably to deter wood predation; and (3) packets of cells of smaller diameter, scattered at various positions in the multiseriate rays. The smaller-diameter ray cells are elongate as seen in a radial section, and can therefore be suspected of having a design more characteristic of flow than storage. Erythrina is a genus of Fabaceae that is typically a tree of seasonally dry subtropical habitats; it may also be a shrub. Most of the species are drought-deciduous. Although the bark of E. caffra is very thin with negligible water content, the wood is mostly devoted to water storage. As can be seen in Fig. 8e, f), the tissue used for this storage is mostly axial parenchyma, not rays. Fibrous bands alternate with prominent bands of axial parenchyma (Fig. 8e). As seen in tangential section (Fig. 8f), the axial parenchyma strands are commonly two cells in length, but some are a single cell in length, and the strands or single cells are storied. The rays of E. caffra (Fig. 8f). The rays are much smaller than those, for example, of M. assurgentiflora (Fig. 8b), and are mostly composed of cells of small diameter as seen in tangential sections. In radial or transverse sections (Fig. 8f), the ray cells are mostly markedly elongate, and can therefore be termed flow cells. Increasing degrees of conversion of secondary xylem to water storage can be seen in the species shown in Fig. 9. Crepidiastrum lanceolatum (Fig. 9a) has upright stems. It is a scarce secondarily woody shrub of the Ogasawara Islands. The secondary xylem consists of vessels, thin-walled ray cells and axial parenchyma, and occasional patches of libriform fibers, which are narrower than the vessels. The libriform fibers are not uniformly distributed, but seem to occur in response to seasonal or other requirement of mechanical strength for upright stems (Carlquist, 1983). Fiber-free zones presumably function in water storage primarily. In Chorisia speciosa (Fig. 9b–d), a tank tree, libriform fibers occur singly or in small groups, which form tangential bands (Fig. 9c). The remaining axial tissue consists of vessels and axial parenchyma. Libriform fibers can be seen in a tangential section (Fig. 9d); the axial parenchyma cells are not markedly elongate, whereas the libriform fibers are very long. The cells of the rays, as seen in tangential section (Fig. 9b) grade from larger cells containing starch to smaller cells relatively poor in starch. The smaller cells are more elongate, as seen in radial section; the starch-storage cells are more nearly isodiametric. In tangential section (Fig. 9d), the axial parenchyma cells grade into the margins of the rays, so that the rays appear poorly delimited. The narrower cells (probably flow cells) are located in the central portions of the rays. Plumeria is a drought-deciduous shrub with thick stems. The secondary xylem (Fig. 9e, f) is very similar to that of Chorisia, although the two are in unrelated families. Alternation of fibrous cells (fiber-tracheids in Plumeria) with axial parenchyma forms an optimal plan for simultaneous strength and water storage in Plumeria. Chorisia,in Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... contrast to Erythrina, has alternation of the two types of cells in bands—amodification of a basic legume wood plan. In tangential section, there is a marked difference, in that Plumeria has well-defined rays composed of radially elongate cells. The tip cells of the rays are vertically elongate, like the axial parenchyma they accompany. The stems of Euphorbia pentagona (Table 1; not illustrated) are quite different, because they have uniseriate rays predominantly and store water in living fibers and axial parenchyma, not unlike other cactiform species of Euphorbia (Carlquist, 1970).

Conversion of Wood (Except for Vessels) Entirely to Storage Parenchyma

The family Caricaceae has woods the rays and background tissue of which are converted to storage parenchyma (Fisher, 1980; Carlquist, 1998). This is illustrated by Jacaratia (Fig. 10a, b). Bands of laticifers (Fig. 10a) are the only exception to that. Rays are composed wholly of upright ray cells, a feature of permanent juvenilism, which is also suggested by the pitting in vessels (Fig. 10b). The remainder of the axial xylem is composed of axial parenchyma strands 1–2 cells in length. Compensating for this minimal-strength configura- tion is the secondary phloem which has prominent bands of phloem fibers (Fig. 10a, top). In fact, xylarium specimens of Caricaceae are sometimes composed wholly of secondary phloem, which is sometimes mistaken for wood because of its fibrous nature. A water storage modality similar to that of Caricaceae occurs in Crassula argentea (Fig. 10c, d). The secondary phloem of C. argentea,however,consistsonlyofthin- walled cells. Rays are very wide (Fig. 10c, d, right halves of photographs). Tannin idioblasts may be found scattered throughout the axial parenchyma and the rays. Axial parenchyma cells are not subdivided into strands (Fig. 10d). Crassula argentea is a succulent in which turgor of water storage cells is responsible for achieving the upright, shrublike form of the plant, which can, in fact, bend over after several months without watering or rainfall. This type of succulence can also be found in the globular cereoid cacti, in which shrinkage between the ribs is prominent, rather than longitudinal shrinkage (Gibson 1970). Kalanchoe beharensis (Fig. 10e, f) and the woodier species of Aeonium,suchas A. arboreum, have an entirely different mode of mechanical support. The wood in these is rayless. A few axial parenchyma cells, insufficient to achieve much water storage, are in contact with the vessels in A. arboreum and K. beharensis. The thick cortex in both of these is the source of water storage, and is without fibers. With respect to wood, most species of Crassulaceae are between the extremes shown for these two species (Metcalfe & Chalk, 1950). Wood of Pereskia aculeata (Cactaceae) is shown in Fig. 11a, b. One would not judge from these photographs that this is the wood of a succulent, and the values in Table 1 show that in fact, it does not rank as a true succulent with respect to wood anatomy. The amount of axial parenchyma does not seem exceptional (Fig. 11a), but as seen in tangential section (Fig. 11b), some of the libriform fibers are wider and thinner- walled (and thereby probably qualify as water-storage cells, whereas others (darker) are narrower and thicker-walled. The amount of ray tissue does seem more than average for a woody species, but not by much. Pereskia aculeata does qualify as a leaf succulent. Cereus repandus (Fig. 11c–e) would qualify as a true succulent based on the much greater proportion of tissue devoted to living cells, especially the rays, and the figures in Table 1 verify this impression. The vessel groupings of this cactus are sheathed by one Author's personal copy

S. Carlquist to three layers of axial parenchyma (Fig. 11c). Interestingly, although axial parenchyma is abundant (Fig. 11e, right), the septa in fibers indicate that the fibers are living also (Fig. 11e. left). In both cereoid and opuntioid cacti, giant primary rays continue without subdivision during secondary growth, so that when thin-walled parenchyma is removed from a stem, a coarse reticulate woody “skeleton” is visible. The rays in Fig. 11d are small by comparison with the giant primary rays which are extended by cambial activity into the secondary xylem. The localities for water storage are quite varied, and may be correlated with growth form and with phylogeny. Some of these are explored by Gibson (1970). A paper by Chapotin et al. (2006) is particularly valuable for showing that in Adansonia, a close relative of Chorisia, water storage works seasonally, rather than daily. We very much need further studies of this kind, but the morphology of succulence and the participation of wood in the water storage program of a plant suggest that succulence usually works on a seasonal basis, with the exception of, for example, halophytes, where salt accumulation via water storage permits growth in saline environments. The plants we typically think of as succulents grow in environments that are highly seasonal. In these environments, such as the summer-wet Colorado Desert of Arizona and adjacent Sonora, or the summer-wet deserts of South Africa, not only is water accumulation accomplished in a few months of the year, flowering and fruiting are also seasonal. Some succulents are notable for their flushing habit, which requires not only water storage, but photosynthate storage as well. Adansonia and Chorisia exhibit flushing growth of shoots, followed in the warmer months by massive flowering and then fruit production. External behavior is thus correlated with the physiological studies of Chapotin et al. (2006).

Photosynthate Storage: What Do We Know?

In the above account, the word “storage” might mean either starch or water storage. Water storage is more readily visible by means of larger cell size, thinner-walled cells, more nearly isodiametric shape (except for living fibers), and greater-than-normal quantity of cells conforming to this description. Photo- synthates, most easily observed in the form of starch, may not be accompanied by any of these special features. For example, in deciduous species of Quercus,ray cells and axial parenchyma cells all contain numerous starch grains during winter. We know that these serve for leafing out, flowering, and fruiting as the year progresses, although we do not have quantitative measures. Starch accumulation and hydrolysis/utilization may occur on a daily basis in actively-growing herbs (Scialdone & Howard, 2015), but deciduous trees, drought-deciduous trees and shrubs, and tank plants must have cycling that features other time periods. We very much need to know the cycling of such events in woody plants in relation to seasonal temperature and water availability. In fact, SEM study of starch grain presence is ideal because it is accomplished simply (liquid-preserved plant portions are required for study), and starch grains can be observed in material fixed at intervals throughout the year. Moreover, the size and surfaces of starch grains can reveal whether they are increasing in size or are eroding away. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Conjunctive Tissue: Origin and Function

The phenomenon of successive cambia was clearly enunciated and correctly applied to a roster of angiosperm species by Pfeiffer (1926). An example is shown here for Stegnosperma (Fig. 11f). Successive vascular increments are produced. Each consists of a vascular cambium that produces secondary phloem and secondary xylem. The vascular increments are produced by a master cambium, a lateral meristem that begins in the cortex and functions for long periods as a single functional meristematic cell layer, producing conjunctive tissue (a type of parenchyma, usually), then a vascular cambium, inwardly (Fig. 11g). The vascular cambium may be dormant after a cycle of initiating a master cambium and conjunctive tissue. This dormancy has led various workers to assume that a vascular cambium, rather than the master cambium in the cortex, is responsible for the process. In plants that grow actively, such as the beet, Beta, the master cambium does not become dormant (until the beet stops growing), but in woodier plants, there can be numerous alternating periods of dormancy and activity in the master cambium, resulting in the corresponding number of vascular increments (Carlquist, 2007b). There have been many erroneous interpretations of the ontogeny of successive cambia, based on illustrations that omit the master cambium or the vascular cambia and thereby do not show the entire process. The interest of successive cambia in the present essay is not so much the ontogeny as what the parenchyma of the conjunctive tissue actually does. In Beta,itstoressugar,but at the same time, it stores ions. The family Chenopodiaceae, to which Beta belongs, is noted for growing in salty soils. In Beta, the conjunctive tissue sequesters salts, thereby being at a kind of osmotic par with salty soil (globular trichomes on the surfaces of Atriplex and Chenopodium leaves are another way of sequestering excess salt). The sugar in the conjunctive tissue forms the basis for the bolting of the single large inflorescence of Beta (Biancardi et al., 2012). Conjunctive tissue in plants with successive cambia is a way of forming parenchyma cylinders alternating with active xylem and phloem, thus providing an excellent way of innervating a storage tissue. In most plants with successive cambia, the vascular cambia remain active for indefinite periods, as evidenced by their continued production of secondary xylem and phloem (Carlquist, 2007b). Conjunctive tissue also serves for enhancing flexibility of lianoid stems, as in Bougainvillea, Boussingaultia, Chamissoa and others. In a sense, the background ground tissue of monocot stems serves for this purpose, as in Dioscoreaceae.

Diversification in Axial Parenchyma Functions

We tend to think of axial parenchyma as anatomically homogeneous, and it often appears to be, but perhaps our observations have been incomplete. This could be caused by the observation of dried wood samples that do not, for example, reveal the occurrence of starch in some axial parenchyma cells of a wood, but absence in others. Ficus elastica (Fig. 12a-e) is introduced as an example here, but there must be many more that are currently unreported. A transverse section of wood of F. ela st ic a (Fig. 12a) shows prominent tangential bands of axial parenchyma, alternating with approximately equally thick bands of libriform fibers. The axial parenchyma (Fig. 12b left) consists of strands mostly 4 cells long. These bands form large contact areas with Author's personal copy

S. Carlquist rays (Fig. 12e). The rays are narrow multiseriate, with mostly procumbent cells but also upright cells contact and simulate the axial parenchyma (Figs. 12b). The tangential bands of axial parenchyma contrast with the sheaths, one to three cells in thickness, of paratracheal axial parenchyma. The paratracheal parenchyma is histologically distinct, and has much smaller cells (compare Fig. 12c, right, with axial parenchyma cells, left). The paratracheal cells are smaller in diameter, Fig. 12d, compared to the cells of the banded axial parenchyma, Fig. 12d, top. The bands of axial parenchyma contain starch (Fig. 12d, top), the paratracheal axial parenchyma cells were not observed to contain starch in the specimen examined. Another type of axial parenchyma cell dimorphism can be found with respect to strands of small crystal-bearing cells (“chambered crystals”). These are reported for caesalpinoid, mimosoid, and papilionate Fabaceae (Metcalfe & Chalk, 1950). “Idio- blastic” axial parenchyma cells are reported in Dinizia (Fabaceae) by Evans et al. (2006), but no difference between these diffuse cells and the paratracheal cells is mentioned. The wood of Robinia (Fabaceae) appears to have diffuse axial parenchyma cells as seen in transverse section (Fig. 12g), but these prove to be merely the widest points of libriform fibers which are identical to other libriform fibers, as can be cofirmed in longitudinal sections.

Fiber Dimorphism as a Source of Living Cells

Fiber dimorphism, monographed recently (Carlquist, 2014) includes instances in which wide living fibers occur in patches whereas other fibers in the same wood are narrower and non-living. This phenomenon is widespread but not common in angiosperms. A previously unreported instance in Eryngium bupleuroides (Apiaceae) is reported here (Fig. 12f). More instances are likely to be reported, and these will demonstrate the flexibility that angiosperms possess for degrees of modification of cell types in order to perform more than a single function.

Functions of Axial Parenchyma in Brassicales: Shifts in Axial Parenchyma During a Year

Marginal axial parenchyma is an umbrella term that includes terminal parenchyma, at the end of a growth ring, and initial axial parenchyma, at the beginning of a growth ring. Transverse sections of wood from four families of Brassicales are included in Fig. 13: Resedaceae (A-C), Brassicaceae (D), and Gyrostemonaceae (E). Reseda alba (Fig. 13a, b), Caylusea hexagyna (Fig. 13c), Stanleya pinnata (Fig. 13d), and Tersonia brevipes (Fig. 13f) are all short-lived perennials or shrubby annuals that grow in dry open areas subject to drying—or an occasional rain—during the growing season. These circumstances may account for the irregular patterning of axial parenchyma. Reseda alba can have initial parenchyma (Fig. 13a)aswellasterminalparenchyma(Fig.13b) in the same stem. Likewise, Caylusea hexagyna (C) can have axial parenchyma bands associated with both wide and narrow vessels. In Stanleya pinnata (D), parenchyma bands extend from latewood into earlywood—perhaps indicating secondary growth during wet winter month, followed by more active growth during warm spring months while the soil is still moist. Tersonia brevipes (E) has small latewood vessels as well as very large earlywood vessels in a background of axial parenchyma, plus diffuse or Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... small groupings of axial parenchyma during the balance of a growth ring. What these three species of woody-herb Brassicales show is sensitive response in vessel diameter and parenchyma presence to availability of water, but also to temperature. The data from comparative anatomy can yield a pattern, but experimental work is much needed to find the physiological explanations for these woods. By determining that, we will learn more about the role of axial parenchyma.

Interxylary Phloem as an Axial Parenchyma Adjunct

Wood of another species of Brassicales, Salvadora persica (Salvadoraceae) is shown (Fig. 13f). This species, like other Salvadoraceae, consists of shrubs or small trees of hot, dry areas. As seen in a transverse section of the wood of Salvadora, vessels are embedded in fibers, and would thus be classified as apotracheal. The axial parenchyma bands are tangential, and of various sizes. In Salvadora, the parenchyma bands increase in tangential length with the diameter of the stem. In these bands, strands of phloem form occasonally. Strands of Interxylary phloem often seem to function as suppliers of carbohydrates during flushes of growth or sudden flowering when water is available (Carlquist, 2013a, b), but the number of species with this peculiar formation is too small to furnish a clear correlation: Interxylary phloem and associated parenchyma can be present in herbs such as Oenothera (Onagraceae) as well as tropical trees such as Strychnos or species of Convolvulaceae.

Living/Septate Fibers as a Dual-Purpose Living Cell Type

Some living fibers do not develop septa, although they remain nucleate for indefinite periods of time (Wolkinger, 1970a, b, 1971). Probably the majority of fibers with extended longevity develop septa; a living cell with such a great length to width ratio can probably function more readily in such respects as starch storage if subdivided into a series of cells. Input and retrieval of photosynthates can be achieved more readily by means of a series of shorter cells, each with its own nucleus and its own pitting, than by means of a single long cell. One can say that all fibers are living at first, but most die upon completing secondary wall formation. Our knowledge of the systematic distribution of living and septate fibers is still limited (Carlquist, 2015a). This can in part be attributed to the preservation of woods in xylaria. Septate fibers can usually be detected in wood samples have been dried, but there may be many more instances of non-septate living fibers that have been missed because living but non-septate fibers cannot be identified in dried wood samples. Presence of septate fibers is often linked to absence or scarcity of axial parenchyma (Carlquist, 2015a). For example, in Rosaceae, axial parenchyma is absent or very scarce in Prunoideae and Spiraeoideae, which have septate fibers. In the other subfamilies of Rosaceae, axial parenchyma is present and septate fibers are not reported (Metcalfe & Chalk, 1950). This is demonstrated in Burseraceae. However, in some genera and species of that family (Trattinckia, Fig. 14a, b), both cell types can be present. The wood of Burseraceae can consist wholly of living cells except for the vessel elements. According to the criteria of Table 1, Burseraceae are subsucculent, which indicates that although the septate fibers serve for water and starch storage, the investment in secondary wall material is sufficient to serve for support in an arborescent growth form. Author's personal copy

S. Carlquist

The same indefinite longevity of fibriform cells can be cited for Nerium oleander of the Apocynaceae (Fig. 15c). This fact, along with other features (stomatal crypts in leaves, vestured pits in vessels) that may explain the remarkable drought tolerance of Nerium. The pervasiveness of living cells in the wood of Apocynaceae may explain how some clades of the family, such as Adenium, have transitioned into succulence. Hedera (Fig. 7f) and other Araliaceae have septate fibers and are notably drought resistant and fire resistant.

Mechanical Significance of Rays

Burgert & Eckstein (2001) and Ozden & Ennos (2014) show that rays are of considerable importance in the tensile and radial stiffness of wood. Ozden & Ennos (2014) indicate that among the woods they studied, Fraxinus has greater resistance to fracture, a fact they attribute to the “homogeneous” (predominance of procumbent ray cells) nature of the rays. The number of species studied in the papers mentioned above is not very great, and we very much need to examine the ray mechanics of other species, which depart further from norms of density in the woods these authors have studied. Some other questions are of major significance: why do rays have an elliptical shape, and why is there a modal distribution of ray height (Metcalfe & Chalk, 1950, Introduction)? Do rays represent weak points in wood if they are vertically more elongate, as in Foeniculum vulgare (Fig. 14d)? Vining and lianoid species of Piper, Aristolochia, Gnetum, and other genera have extremely tall rays, which can be mea- sured in centimeters rather than microns. Do such rays confer greater flexibility that permits vines and lianas to twist in relation to plants that support them? The significance of very wide rays may have more than one explanation. In Betulaceae and Fagaceae, aggregate rays are formed as a result of coalescence of uniseriate rays. This may have both physiological and mechanical explanations.

Rayless Woods Can Have Living Cells

Raylessness can occur in the first growth ring of some genera, such as Artemisia.This has been called early onset raylessness (Carlquist, 2015b). Rays arrive at some point later. This also happens in the insular species of Plantago. Lack of living cells in the early-formed secondary xylem of these genera is understandable because the value of rays and axial parenchyma becomes greater with increase in stem diameter, judging from anatomical data. The presence of axial parenchyma in a rayless wood need not be accompanied with presence of rays. In Pimpinella dendrotragium (Fig. 14e-f). Only two very small rays were observed in the wood of my sample of P. dendrotragium (Fig. 14g). However, all vessels in this species are accompanied by axial parenchyma cells, which may even be longitudinally subdivided (Fig. 14f). In Kalanchoe beharensis (Fig. 10e, f)woodis entirely rayless but there are axial parenchyma cells surrounding the vessels. Septate and living fibers are present in the wood of P. dendrotragium and may account for radial transportation of solutes. In a rayless species, axial parenchyma may be more important than rays if it has a physiological value in osmotic maintenance of water columns. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

The significance of rayless woods appears to be a temporary increase in some kinds of mechanical strength at the expense of radial transport capabilities (Carlquist, 2015b), a balance that can be reversed if a stem increases in diameter and ultimately develops rays. Stems that increase in diameter but never form rays may manage radial transport by means of living fibers. The wood of the species of Hebe (Veronica) that I have examined consists of vessels, living (but non-septate) fibers, and vasicentric tracheids.

Axial Parenchyma Can Have Multiple Functions in a Given Stem at Different Times

An obvious but infrequently mentioned function of axial parenchyma in stems and roots is its relationship to protoxylem. Protoxylem vessel elements and tracheids are extensible, and thereby require contact with equally extensible parenchyma cells (Fig. 15a, g). The illustrations to demonstrate this are taken from Araliaceae and its sister family, Apiaceae. In a transverse stem section of Foeniculum vulgare (Fig. 15a) andinaradialsectionofOreopanax steinbachianus (Fig. 15g) clear sequences from protoxylem to metaxylem can be seen. As seen in transverse section, the axial parenchyma cells that surround protoxylem vessels expand to extinguish canals left by collapse of annular and helical vessels (Fig. 15a). This collapse is not so evident in Fig. 15g, left. One can question whether one should refer to parenchyma around protoxylem vessels as axial parenchyma, but it does form cylinders that surround the metaxylem vessels (Fig. 15a,top,Fig.15f,left)and early secondary xylem vessels (Fig. 15b, center, g, right). A minimum of starch can be seen in the parenchyma of the protoxylem (Fig. 15g, left) and in the metaxylem (Fig. 15g, center). Just to the right of the metaxylem vessel in Fig. 15g is a septate fiber that contains abundant starch. We very much need studies that capture the deposition and removal of starch from these various sites, because differential activity is very likely to be found. SEM is a promising method to explore this. The axial parenchyma that surrounds metaxylem and early secondary xylem (Fig. 15g, center and right; Fig. 16f) is not involved in elongation, but rather probably functions in osmotic maintenance of the conductive stream. The vessel elements shown in Fig. 16f are all pitted metaxylem vessels, but the near-simple perforation plate (Fig. 16f, right) indicates a transition to secondary xylem. This pattern of axial parenchyma sheathing of vessels continues until the end of the growing season, which ends in serious drought for Foeniculum vulgare. The stem studied lived for two years; the very narrow vessels (and very likely some vasicentric tracheids) can be seen in Fig. 15c (arrows). A radial section of the end of the second year’s growth shows an even more pronounced narrowing of the vessels, which are very numerous, accompanied by more parenchyma. Some of the vessels are so narrow that they are imperforate (Fig. 15e), and thus qualify as vascular tracheids. Narrower vessels and vascular tracheids resist embolization to a greater extent than wider vessels (Hargrave et al. 1994). At the end of the growth of an annual or short-lived perennial that terminates as soil reaches drought conditions, axial parenchyma seems exceptional in quantity and also perhaps in function. If so, it could be considered a kind of drought axial parenchyma, supplying water to seeds as the plant ceases to grow. Author's personal copy

S. Carlquist

Cell Type Diversification Within Rays

Multiseriate rays of Araliaceae contain procumbent cells predominantly (Fig. 16e). This is also true in a sister family of Araliaceae, Myodocarpaceae (Fig. 17a). The sections of Oreopanax wood (Fig. 16a–d, f) were made from a young stem with relatively little secondary xylem, because these relatively juvenile stems have radially shorter ray parenchyma cells, and thus one can encounter end (tangential) walls more readily than in older stems (Fig. 16e). Light microscope examination of tangential walls of ray cells shows that they are pitted, but SEM is desirable to show details of this pitting. In Fig. 16a (left) we see an end wall exposed, with bordered pits on the outer surface. A view of the inner surface of an end wall can also be seen (Fig. 16a, right). An end wall of a narrower ray cell (Fig. 16b) shows that large pits are present, with intervening strands of wall material. This is shown to a greater degree in Fig. 16c, right, where pits occupy most of the wall. Such cells can be considered flow cell. This is also suggested by the absence of starch grains in the two cells designated as flow cells here. The cell to the left in Fig. 16c can be considered a storage cell, because the pits are small and sparse, and starch is present. There is the possibility that starch grains can be displaced during the sectioning process, so several similar cells should be examined before this conclusion is reached. The tip cell of a ray is also presented (Fig. 16d). At the top, starch grains are seen, but below, a view of the outer surface of a tangential wall of this tip cell is evident. The pits are moderately dense, but small and bordered. This tip cell is probably functionally a storage cell. All of these pits in Fig. 16a–d are probably bordered—no simple pits were observed on tangential walls of Oreopanax in this survey. The pits on radial walls of the procumbent ray cells of Oreopanax are very small and very sparse (Fig. 16e). These pits contact either other ray cells or septate fibers. Conduction across these interfaces must be minimal. The occurrence of tile cells in rays of Malvales, monographed by Chattaway (1933) is an example of diversification of ray cells, because tile cells represent axial subdivi- sions of procumbent cells, and could thus be considered an additional type of ray cell. We do not know the function of tile cells, which are apparently restricted to only some genera of Malvales. This is an example of how problematic it may be to study the function of particular parenchyma cell manifestations in wood.

Juvenilism in Wood: Redirection of Flow Patterns of Photosynthates

In my paper on the occurrence of protracted juvenilism (paedomorphosis; neoteny of some authors), I described the patterns seen in rosette trees, rosette shrubs, succulents, and various annuals and perennials (Carlquist, 1962). In these species, there is a descending curve for length of vessel elements, and a corresponding delay of circular pitting in vessels; scalariform and pseudoscalariform pitting, such as is seen in Fig. 16f (top center) continues indefinitely into the secondary xylem. In that paper, I also mentioned the tendency for upright cells in rays, common at the beginning of secondary growth in most plants, to be produced indefinitely in the secondary xylem. Such rays (e.g., Fig. 17f) were integrated into the Kribs (1935) system of ray types (Carlquist, 1988,200–204) as “Paedomorphic Type I,” etc. Raylessness was considered a kind of ultimate expression of juvenilism: the delay of ray production, and the production of Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the... upright cells that were so much like libriform fibers that they were indistinguishable from them. The anatomical data presented in those sources seem entirely accurate, but no physiological explanations for these conditions were attempted at that time. Given the concept that flow in elongate cells occurs in the direction of cell elongation and the designation of radially elongate ray cells as flow cells in the present paper, an interpretation of juvenilism in rays that hypothesizes patterns of flow vertically rather than horizontally seems justifiable. One must be careful in reporting cell shape of rays because particular rays may not be sectioned through their central portions (sagittally), and thus fewer procumbent ray cells may appear to be present. In the woody families of Apiales, Myodocarpaceae (Fig. 17a) and Araliaceae, ray cells in multiseriate rays are predominantly procumbent. This may related to the production of rather large leaves in these families, as well as other factors, such as growth in flushes. These two families are somewhat unusual in this respect, because the bulk of the families of woody angiosperms have multiseriate rays that are typically heterogeneous or heterocellular: composed of cells upright, square, and procumbent in radial sections. With woody Apiales as a basal type in the order, juvenilism in rays would be expected to be characterized by fewer procumbent cells as well as the square and upright cells characteristic of paedomorphic rays in angiosperms generally. Apiaceae can be considered a predominantly herbaceous derivative of Araliaceae in which instances of secondary woodiness, which would involve paedomorphosis (Carlquist, 2009), occur. If we examine wood of Apiaceae (Fig. 17b–e)wefindthat this is true. Both Heteromorpha trifoliata (Fig. 17b)andEryngium inaccessum (Fig. 17c) have wood in which a few procumbent cells are present, but most ray cells are square or upright. Heteromorpha trifoliata is an African small tree from highland elevations, where the climate is close to temperate throughout the year. Eryngium inaccessum is a shrubby species, on the moist and temperate Juan Fernandez Islands, of a genus that is otherwise mostly rosette-forming. Arracacia atropurpurea (Fig. 17d– e) is a somewhat woody species in a genus that contains carrot-like herbaceous species. It grows on temperate Mexican uplands. In A. atropurpurea, wood begins with rays composed wholly of upright cells (Fig. 17d). Over time, as the plant becomes woodier, some procumbent cells are produced (Fig. 17e). All three of these species grow in conditions that do not change much throughout the year. In these climates, secondary woodiness would be likely to occur. Secondary woodiness is an economical form of increasing plant mass if conduction can be sustained throughout the year. Axial parenchyma may provide vertical flow of photosynthates to sustain growth. In annuals and short-lived perennials, upright axial parenchyma cells may convey photosynthates for flowering and fruiting. Uniseriate rays and uniseriate wings of multiseriate rays may offer links for vertical conduction of photosynthates from rays into axial parenchyma. Juvenilistic rays can be found in a series of angiosperms that are not rosette trees or shrubs, nor are they annuals or perennials or succulents, which typically have juvenilistic wood. Chloranthaceae is a family with many “primitive” characters that has protracted wood juvenilism. Rays in this family are mostly composed of upright cells, as in Hedyosmum (Fig. 17f). Hedyosmum species can be trees, but not large ones, and the overall habit in the genus is shrubby. Chloranthus has even more pronounced ray cells (Fig. 17g), which are composed entirely of upright ray cells. Chloranthus consists of subshrubs branched from the base. The habit of protracted production of upright cells may be found in some Eupomatiaceae, Winteraceae, Aristolochiaceae, Author's personal copy

S. Carlquist

Piperaceae and Austrobaileyaceae. Individuals of some species in these families develop from shrubs into trees, at which juncture procumbent cells become more abundant. The occurrence of radially-elongate (procumbent) cells in rays is certainly related to arborescence. The bigger the diameter of a stem, the greater the value of radial flow of photosynthates into and out of the stem. We should not be surprised that as Barghoorn (1941) showed, rays become wider and contain more procumbent cells in wider stems. The fact that rays also subdivide suggests that mechanical optima are being served. The poles of herbaceousness and woodiness were not appreciated by earlier wood anato- mists, who wanted to study “mature” wood patterns of species that were often commercially valuable timbers. The idea that ontogenetic changes in ray, axial parenchyma, and tracheary elements (fibriform cells, including tracheids; vessel elements) can be shifted toward herbaceous modes or woodier modes (Carlquist, 1962, 2009; Carlquist, 2013a, b) places us in a new perspective. Features of protracted juvenilism on the one hand, or accelerated adulthood on the other, can be selected to achieve translocation of photosynthates and other substances vertically or horizontally. Like- wise, the mechanical properties of earlier-formed secondary xylem can be quite different from wood of an older stem. Axial parenchyma and ray parenchyma are important elements in these ontogenetic changes. The success of angiosperms in no small measure derives from the fact that degrees and kinds of shifts from juvenile to adult patterns of wood structure can be increased or decreased, a flexibility not possible in conifers, Gnetales, cycads, or Ginkgo.

Conclusions

The following conclusions can be derived from the studies reported above, combined with findings in literature cited in this paper. Some of the statements represent hypotheses that seem supported by wood anatomy, but additional evidence is needed on a number of points.

1. Anatomical features of living wood cells should be regarded as indictors of probable function. We have so many ways of determining activities of these cells that retreat to a purely descriptive approach, guided by what is in glossaries, should be regarded as outmoded methodology. When we view the variability in anatomical expressions in these cells within a plant and among species, we should be prepared to see that they represent optimal structural modes, not relictual details that have failed to change. 2. Gnetales are a vessel-bearing group that we now know are part of the conifer clade and not a derivative or ancestor of angiosperms. Because Gnetales have acquired vessels independently of angiosperms, they show us the anatomical and physiological consequences of vessel presence. These include the function of multiseriate rays and axial parenchyma (living fiber-tracheids in the case of Ephedra)to conduct photosynthates into the wood where they can participate in functions of storage and relate to the conductive process. Multiseriate rays can function in providing flexibility of stems (hence the numerous lianoid species of Gnetum; some species of Ephedra are also lianoid). Gnetales have essentially all of the xylary features of woody angiosperms, and therefore represent a kind of Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

experimental verification of how equivalent angiosperm structures function. An- giosperms have exceeded Gnetales in evolutionary diversification mostly because the long gymnospermous life cycle of Gnetales makes for slow reproduction and slow territorial expansion of species. Gnetales have conducting systems that seem quite equivalent to those of angiosperms in functional characteristics. 3. Rays and axial parenchyma in angiosperms form a continuous network; axial parenchyma cells cannot exist isolated from other living cells. Rays are the primary point of entry for carbohydrates into the secondary xylem, and through contacts with axial parenchyma provide entry of photosynthates and ions to axial parenchyma strands. 4. Although most conifers have axial parenchyma strands, some (certain species of Agathis and Araucaria; some species of Picea and Tsuga)donot.Thisisan indication of a subsidiary role for axial parenchyma in conifers compared to angiosperms (and Gnetales). Axial parenchyma in conifers is part of a syndrome of differences in hydraulic systems between conifers and angiosperms. Tracheids in conifers are sufficiently long that all tracheids contact one or more rays. Rays in conifers are almost never more than one cell layer wide, and diversification in function is thereby limited (except for presence of ray tracheids, primarily in Pinus). Presence of prominent tracheid-to-ray pitting is indicative of a role for rays in conduction in tracheids. 5. The tracheids of conifers are narrow enough so that water columns are easily restored when water frozen in them thaws. Wood of tropical conifers in localities where water in tracheids is not likely to freeze can exceed that limitation in diameter. The diameter of circular bordered pits in conifer tracheids (which is less than tracheid diameter), restricts the amount of the water-conducting margo porosities. The conductive capability of the pit aperture, and the ability of pits to aspirate form a syndrome of safety for conifers that represents a pattern different from that of ray plus axial parenchyma in angiosperms. With functions of axial parenchyma and rays in conifers limited (or adequately served by fewer cells), volume occupancy by greater volume of tracheids and smaller volume of living cells becomes a conifer wood strategy. 6. Axial parenchyma in angiosperm woods serves for maintenance (by osmotic functioning) to deter and repair embolisms. Ray parenchyma introduces photosynthates to the parenchyma network and thereby makes this possible. The function of parenchyma in maintaining flow in vessels and tracheids of angiosperm woods provides a pre-existing system that can take on other functions: water storage, photosynthate storage, defense against predation, zones of flexibility, etc. 7. Axial parenchyma and (indirectly) rays, vessel grouping, narrowing of vessels, and presence of tracheids and other mechanisms provide conductive safety in angiosperms and supersede scalariform perforation plates. Scalariform perforation plates offer devices for potentially confining air bubbles, but have the disadvantage of providing a high degree of resistance to flow. This resistance is tolerable in plants with low and steady rates for flow. Such plants are either in mesic habitats, such as cloud forests, or have microphyllous foliage (Bruniaceae) to minimize transpiration. Reduction in number and thickness of the bars in perforation plates can be seen in some families (Hamamelidaceae, Monimiaceae) as ways of Author's personal copy

S. Carlquist

reducing flow resistance, but these midpoints on the way to simple perforation plates are relatively few. 8. Axial parenchyma is sparse and often diffuse in “primitive” angiosperm woods, as it is in many conifers. With the development of vessels, axial (and indirectly, ray) parenchyma become sources of osmotic support of the water columns of vessels. Accordingly, volume devoted to axial and ray parenchyma increases markedly in vessel-bearing angiosperms, and this increased volume parallels degrees and kinds of increase in vessel presence, vessel diameter, and vessel area as seen in transverse sections. Similar trends may be cited for the increase in axial and ray parenchyma in Gnetales as compared to conifers. 9. In angiosperms, conversion of axial and ray parenchyma to sites for deposition of secondary compounds for deterrents of phytophagous insects and fungi is generally slower and more partial than it is in conifers. Angiosperm rays (and to a lesser extent axial parenchyma) show marked division of labor into flow cells, storage cells, and “defense” cells. Amorphous deposits of tannins, resin-like compounds, and latex, as well as crystalline (usually calcium oxalate crystals of various sizes and forms) and non-crystalline dissolved compounds serve for defense purposes. Over time, some of these substances may extend into storage and flow cells (as well as vessels and tracheids) in portions of secondary xylem that are no longer active. 10. Elongate shape of ray and axial parenchyma cells is indicative of direction of flow through those cells. End walls (tangential walls of ray cells, horizontal walls of axial parenchyma strands) show pits that are wider, more densely placed, more numerous, and are frequently bordered. Presence of borders allows for more flow through wider pit membranes plus minimal interruption of the secondary wall. Denser, larger, and more numerous pits also increase flow possibilities. Pitting on lateral walls of elongate cells and on storage cells, which are more nearly isodiametric, is sparser and consists of smaller pits. The secondary walls of ray and axial parenchyma cells are sometimes described as “nodular” or “nodulated,” but this is a misnomer for the borders of pits or the thickness of secondary walls between pits on ray or axial parenchyma walls. 11. Walls of phloem ray parenchyma and axial phloem parenchyma cells are usually thin, and often lack secondary walls. Such walls can be considered equivalent to use of the entire cell wall as a pit membrane, and thus valuable for flow of photosynthates. Although the pathways of sugars from phloem into secondary xylem have not been studied sufficiently, the phloem and xylem rays represent, by default, the most obvious and potentially most efficient route for this flow. 12. Vessel-to-ray and vessel-to-axial parenchyma pit areas are often prominent in area and density. This is also true of the contact points between rays and tracheids in conifers, producing the various patterns collectively known as “cross-field” pitting. The large areas of pit membranes in these contact points in gymnosperms and angiosperms, respectively, suggest a correlation with release of sugars and ions into the conductive stream. Although the relative paucity or absence of axial parenchyma in conifers may suggest a smaller role in conifers for such osmotic action, the great length of tracheids in most conifers has the effect of putting all tracheids in contact with at least one ray. In angiosperms, a similar statement could be mode, not only with regard to vessel-ray contacts and vessel-axial parenchyma contacts, but also axial parenchyma-ray parenchyma contacts (the “contact cells” of some authors). Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

13. Photosynthate storage (mostly as starch) and water storage in ray and axial parenchyma of angiosperms permit development of growth forms not possible in conifers: succulents, lianas, rosette trees, trees that grow by means of flushing, drought-deciduous trees, tank trees, and many others. This diversity of growth forms in angiosperms can be considered as made possible by the devotion of more volume to water and starch storage cells. Thus, the introduction of parenchyma volume (perhaps primarily as a means of osmotic regulation of flow in vessels) in earlier angiosperms has formed the basis for this radiation. 14. The strategies of water storage within secondary xylem are not uniform. Storage in rays may be more prominent in some angiosperms, storage in axial parenchyma more prominent in others. In some plants, the entirely of the secondary xylem other than vessels is devoted to water storage and/or starch storage. Bark may be the primary source of water and/or starch storage in some species. Turgor can have mechanical aspects as well as being related to water storage. 15. Water storage cells in axial parenchyma and rays tend to be larger, more sparsely pitted, capable of expansion and contraction, and without visible contents as compared to parenchyma cells that serve for photosynthate storage or other purposes (accumulation of defensive substances). Liquid-preserved wood samples are required to establish clearly the function of living cells. Function of wood cells can be inferred from study of dried wood samples, but not determined with certainty. The volume of starch in axial parenchyma cells and ray cells is considerable, and should be figured, wherever possible, in wood studies (especially those designed for students). Although we have no good comparative data, workers familiar with study of liquid-preserved wood samples occasionally remark that the totality of starch storage is perhaps never mobilized, and thus a “standby” supply is present. 16. Living fibers (which may be septate or non-septate) are part of the system of living cells in certain woods, and can serve for storage (usually starch). Living fibers, where abundant, may be accompanied by reduction in abundance or presence of axial parenchyma, although exceptions exist. Living fibers represent a compromise between mechanical strength (by virtue of greater wall thickness and fiber length) and photosynthate storage capabilities (wider lumen diameter). 17. Rays have mechanical strength properties that can be analyzed in terms of various forces (stiffness, torsion, etc.), and these factors are definitely part of ray design in any given species. Only a small number of woody species have been studied in this regard, and many more studies are needed to develop a picture. As seen in tangential section, rays have an ellipsoidal shape and are more commonly finite in height (less than about 500 μm in vertical length). Rays vertically longer than this may be associated with greater stem flexibility or other properties, and are thus valuable for vines, lianas, and some kinds of herbs and subshrubs. Ellipsoidal shape of rays would tend to interrupt the strength pattern conferred by fibers less than rays of indefinite length; the diagonal structural members interconnecting linear bridge parts is a parallel. 18. Ray cells in conifers, as a generalization, have thinner walls than those of angiosperms. The uniseriate nature of rays in conifer woods suggests that a smaller volume devoted to parenchyma is correlated with a greater volume devoted to tracheids, and that storage of either water or photosynthates in rays Author's personal copy

S. Carlquist

of conifers is much less that in angiosperms. The radially elongate nature of conifer ray cells also gives them the aspect of “flow” cells rather than “storage” cells. Rays in conifers also interrupt the mechanical strength pattern of conifer woods minimally. In angiosperms, rays are wider, taller, and greater in total volume. The gain in volume for photosynthate and water storage has the potential of lessening the tensile and torsion aspects of mechanical strength, but this can be countered by greater thickness of secondary walls in ray cells and greater thickness of fibriform wood cells (libriform fibers, fiber-tracheids). The greater width of rays in angiosperms as compared to those of conifers offers potential advantages for growth forms that experience more torsion, such as lianas. 19. Many angiosperm woods have limited numbers of axial parenchyma cells associated with vessels or vessel groupings (vasicentric or paratracheal scanty are terms applied). This suggests that the function of osmotic control of conduction in the vessels is accomplished with a limited number of cells. Larger aggregations of axial parenchyma may therefore be involved in part in some other function as well. Multiplicity of functions is sometimes indicated by differing cell morphology (e.g., the two types of axial parenchyma in Ficus; crystal strands in axial parenchyma of Fabaceae) but is usually not apparent. 20. Angiosperm axial parenchyma in transverse section is often labeled as either apotracheal (no consistent association pattern of vessels with axial parenchyma) or paratracheal (vessels always associated with parenchyma in variously shaped or positioned aggregations). If we view these two types in three dimensions, we find both that vessels shift in position along their vertical course, and all vessels contact axial parenchyma (and rays) somewhere along their length. No clear correlations of apotracheal and paratracheal arrangements have been demonstrated. The causes for divergent modes of axial parenchyma distribution should be investigated. 21. Axial parenchyma and rays may shift in abundance and wall characteristics, as functions in a stem change. For example, axial parenchyma and rays are minimal in quantity and thick-walled early in secondary growth in Ipomoea (and other vines and lianas), but abundant and thin-walled in later-formed wood, as there is a shift from upward progress to flexibility in stems. Other examples include shifts from fibrous (stiffness) to parenchymatous (water storage). 22. Some investigators have pointed to correlations between particular character states in wood. For example, diffuse axial parenchyma is often claimed to be associated with heterocellular rays, presence of tracheids as the imperforate tracheary type, and vessels with scalariform perforations plates. The exceptions to this correlation and other correlations are numerous. We may attribute that to homoplasies, or to the idea that optimal wood patterning may be diverse because cell types are not uniform (e.g., rays are polymorphic in cell types and in the proportion of the various types; in dimensions; and in numbers per horizontal mm). Attempts have been made by some workers to refer particular wood types to “stages” of organization. Rather than begin with an attempt to refer a given wood to particular types or correlations, analyzing woods on an individual basis with relation to growth form, ecology, and other factors is likely to be more productive to our understanding of how woods function and evolve. 23. Upright cells (as seen in radial sections) in rays are more likely to serve for vertical transportation of solutes, whereas procumbent cells architecturally seem ideal for Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

horizontal photosynthate flow. Upright cells, especially where abundant, are likely to intersect with axial parenchyma more often, and thereby form links between horizontal flow and vertical flow. Square cells have been claimed to be “morphologically equivalent to upright cells,” but they are better regarded as cells that serve functions related to isodiametric shape (e.g., storage or interconnection between axial and radial cells) better. Density and number of pits are probably excellent clues to function in ray cells. These features can best be studied with SEM. 24. Protracted juvenilism (heterochrony, paedomorphosis, or neoteny of various authors) is a feature of angiosperm woods almost exclusively. Where ray parenchyma is concerned, juvenilism means the production of upright cells exclusively or mostly so, sometimes for the entire life of the plant. This has the effect of directing photosynthate flow vertically. Vertical flow of water containing sugars in parenchyma of woods is easily understood in rosette trees and rosette shrubs, in which diameters of stems do not increase very much over time. Most trees have marked increase in stem (and root) diameter, and therefore feature radial flow, to supply photosynthates to storage sites and to the axial parenchyma. Radial flow is achieved mostly in procumbent cells, which become more numerous as a tree increases in diameter. Vertical flow in upright ray cells would correlate with flow of photosynthates in the wood to flowers and fruits that terminate shoots in annuals and perennials (“woody herbs”). Vertically longer rays are often a manifestation of juvenilism, because in woody species they are not subdivided ontogenetically over time. Protracted juvenilism has other manifestations in vessel-element length, vessel pitting, and mechanical cell length. By being an “ontogenetic intervention” (as opposed to accelerated adulthood, seen uniformly in conifer woods), the ability to protract juvenilism is a rich source of architectural repatterning present in angiosperm woods. Protracted juvenilism does not occur merely in the wood of a few woody herbs on islands; it occurs in annuals and perennials throughout the world. 25. Raylessness can be considered a penultimate form of juvenilism, in which upright ray cells are produced exclusively and simulate fibers so closely that they are indistinguishable. Some rayless woods develop rays eventually, demonstrating that the adult condition of having rays has been delayed. Raylessness in this case may be considered an initial emphasis on mechanical strength of fibriform cells over parenchyma cells designed for radial flow, followed by a balance between the two. Rayless woods are not all alike; some may have living fibriform cells, some have axial parenchyma but no rays. The ultimate form of juvenilism is represented by loss of the vascular cambium, which has occurred in monocots. Monocot stems counteract the loss of a vascular cambium by producing vascular tissue in a wider cylinder of scattered bundles embedded within a parenchymatous (or sclerenchy- matous) background. Some shrubby, rosette-forming (Agave), or arborescent monocots (Dracaena) have restored the value of secondary growth via a monocot cambium which produces bundles and associated parenchyma inwardly. 26. Plants with successive cambia, such as Beta, the beet, produce rings of vascular tissue, each with secondary phloem, a cambium, and secondary xylem, separated from each other by conjunctive tissue. In the majority of woods with successive cambia, conjunctive tissue is composed of parenchyma cells. Conjunctive tissue is quite different from axial parenchyma: most plants with successive cambia have axial parenchyma in the secondary xylem, just as plants with a single Author's personal copy

S. Carlquist

cambium do. The function of conjunctive tissue may be in storing sugars (Beta) or starch; in sequestering salts; in achieving flexibility (lianas); or in providing living phloem and xylem alternating with tissue devoted to storage, thereby shortening input and retrieval pathways. 27. The size, location, density, location, number and morphology (e.g., presence of borders) of pits on parenchyma cells in woods are strongly correlated with function. These features can be seen in sectional view. However, they are probably best displayed when we look at the outer surfaces of parenchyma cells. Preparations made by hand sectioning with single-edge razor blades and examined with SEM are particularly valuable, because cells are often separated rather than sliced open, so cell contexts, cell shapes and details of pitting are revealed. We lose information when we limit ourselves to a single method of preparation. Wood is a three-dimensional structure, and if sections are convenient ways of obtaining information (because they present easily-grasped two-dimensional images of wood), we must conceptually reconstitute the three-dimensionality of woods if we are to understand them thoroughly. 28. Ray parenchyma and axial parenchyma configurations, like other features in wood, should be regarded not as historical markers of levels of “specialization” achieved by lineages fortunate enough to escape “primitive” or “unstable” conditions. Rather, character states in woods living today are likely to be optimal in particular localities in particular growth forms. Some of these optimal construc- tion plans in wood may contain more “primitive” features than others, but by being alive today, they are all still functional. The fact that early angiosperms probably had wood with multiseriate rays and some degree of axial parenchyma means that diversification in these tissues could be utilized as a source for development of growth forms and adaptations that have not be achieved by conifers and other vascular plants. Individual species quite frequently diverge from generalities derived from quantification of large numbers of species. These apparent deviations from mathematical norms, averages, or modes may tell us more about adaptation than the data points that adhere more closely to mathematically-obtained curves for groups of species.

Literature Cited

Bailey, I. W. 1944. The development of vessels in angiosperms in morphological research. American Journal of Botany 32: 421–428. Barghoorn E. S. 1941. The ontogenetic development and phylogenetic specialization of rays in the xylem of dicotyledons. 2. Modification of the multiseriate and uniseriate rays. American Journal of Botany 28: 273–282. Biancardi, E., L. W. Panella, & R. T. Lewellen. 2012. Beta maritima. The origin of beets. Springer Verlag, Heidelberg. Bowe, L. M., G. Coat & C.W. dePamphilis. 2000. Phylogeny of seed plants based on all three genomic compartments: extant gymnosperms are monophyletic and Gnetales’ closest relatives are conifers. Proceedings of the National Academy of Sciences USA 97: 4092–4097. Braun, H. J. 1970. Funktionelle Histologie der sekundären Sprossachse. I. Das Holz. Encyclopedia of Plant Anatomy IC, part 1. Borntraeger, Berlin. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Braun, H. J. 1984. The significance of the accessory tissues of the hydrosystem for osmotic water shifting as the second principle of water ascent, with some thoughts concerning the evolution of trees. IAWA Bull. n.s. 5: 275–294. Brodersen, C. R., Knipfer, T. & A. J. McElrone. 2017. In vivo visualization of the final stages of xylem refilling in grapevine (Vitis vinifera) stems. New Phytologist 217: 117–126. Burgert, I. & D. Eckstein. 2001. The tensile strength of isolated wood rays of beech (Fagus sylvatica L.) and its significance for the biomechanics of living trees. Trees 15: 168–170. Burgert, I., A. Bernasconi & D. Eckstein. 1999. Evidence for the strength function of rays in living trees. Holz als Roh- und Werkstoff 57: 397–399. Burleigh, J. G. & Mathews. 2004. Phylogenetic signal in nucleotide data from seed plants: implications for resolving the seed plant tree of life. American Journal of Botany 91: 1599–1613. Carlquist, S. 1962. A theory of paedomorphosis in dicotyledonous woods. Phytomorphology 12: 30–45. Carlquist, S. 1970. Wood anatomy of Hawaiian, Macaronesian, and other species of Euphorbia.Botanical Journal of the Linnean Society 63 (Supplement 1): 181–193. Carlquist, S. 1975. Ecological strategies of xylem evolution. University of California Press, Berkeley. Carlquist, S. 1983. Observations on the vegetative anatomy of Crepidiastrum and Dendrocacalia. Aliso 10: 383–395. Carlquist, S. 1988. Comparative wood anatomy. Springer Verlag, Heidelberg. Carlquist, S., & D. A. Gowans. 1995. Secondary growth and wood histology of Welwitschia. Botanical Journal of the Linnean Society 118: 107–121. Carlquist, S. 1998. Wood and bark anatomy of Caricaceae: correlations with systematics and habit. IAWA Journal 19: 191–206. Carlquist, S. 2007a. Bordered pits in ray cells and axial parenchyma: the histology of conduction, storage, and strength in living wood cells. Botanical Journal of the Linnean Society 153: 157–168, Carlquist, S. 2007b. Successive cambia revisited: ontogeny, histology, diversity, and functional significance. Journal of the Torrey Botanical Society 134: 301–332. Carlquist, S. 2009. Xylem heterochrony: an unappreciated key to angiosperm origins and diversification. Botanical Journal of the Linnean Society 161: 26–65. Carlquist, S. 2012a. How wood evolves: a new synthesis. Botany 90: 901–940. Carlquist, S. 2012b. Wood adaptation of Gnetales in a functional, ecological, and evolutionary context. Aliso 30: 33–47. Carlquist, S. 2013a. Interxylary phloem: diversity and functions. Brittonia 65: 477–495. Carlquist, S. 2013b. More woodiness/less woodiness: evolutionary avenues, ontogenetic mechanisms. International Journal of Plant Sciences 174: 964–991. Carlquist, S. 2014. Fibre dimorphism: cell type diversification as an evolutionary strategy in angiosperm woods. Botanical Journal of the Linnean Society 174: 44–67. Carlquist, S. 2015a. Living cells in wood. 1. Absence, scarcity and histology of axial parenchyma as keys to function. Botanical Journal of the Linnean Society 177: 291–321. Carlquist, S. 2015b. Living cells in wood. 2. Raylessness: histology and evolutionary significance. Botanical Journal of the Linnean Society. 178: 529–555. Carlquist, S. 2017. Conifer tracheids resolve conflicting structural requirements: data, hypotheses, questions. Journal of the Botanical Research Institute of Texas 11: 123–141. Chapotin, S. M, J. H. Razanameharizaka & N. M. Holbrook. 2006. Water relations of baobab trees (Adansonia spp. L.) during the rainy season: does stem water buffer daily water deficits? Plant Cell and environment 29: 1021–1032. Chattaway, M. M. 1933. Tile cells in the rays of Malvales. New Phytologist 32: 261–273. Davis,S.D.,J.S.Sperry,andU.G.Hacke.1999. The relationship between xylem conduit diameter and cavitation caused by freezing. American Journal of Botany 86: 1367–1372. Evans, J. A., P. E. Gasson & Gwilym P. Lewis. 2006. Wood anatomy of the Mimosoideae (Leguminosae), IAWA Journal, Supplement 5: 1–117. Fisher,J.B.1980. The vegetative and reproductive structure of papaya (Carica papaya) Lyonia 1: 191–208. Gibson, A. C. 1970. Comparative anatomy of secondary xylem in Cactoideae (Cactaceae). Biotropica 5: 29– 65. Greguss, P. 1955. Identification of living gymnosperms on the basis of xylotomy. Akademiai Kiado, Budapest. Greguss, P. 1959. Holzanatomie der Europaïschen Laubhölzer und Sträuchern. Akademiae Kiado, Budapest. Greguss, P. 1968. Xylotomy of the living cycads. Akademiai Kiado, Budapest. Author's personal copy

S. Carlquist

Hacke, U., B. G. Lachenbruch, J. Pitterman, S. Mayr, J.-C. Domec & P. Schulte. 2015. The hydraulic architecture of conifers. In, U. G. Hacke, ed. Functional and ecological xylem anatomy, 39–75. Springer Verlag, Heidelberg. Hargrave, K. R., Kolb, K. J., Ewers, F. J. & S. D. Davis. 1994. Conduit diameter and drought-induced embolism in Salvia mellifera Greene (Labiatae). New Phytologist 126: 695–705. Heady,R.D.,J.G.Banks&P.D.Evans.2002. Wood anatomy of Wollemi pine (Wollemia nobilis Araucariaceae). IAWA Journal 23: 339–357. Hoadley 1990. Identifying woods: accurate results with simple tools. Taunton Press, USA. Holbrook,N.M.,&M.A.Zwieniecki.1999. Embolism repair and xylen tension: do we need a miracle? Plant Physiology 120: 7–10. Holbrook, N. M., Zwieniecki, M. A., & P. J. Melcher. 2002. The dynamics of dead wood: maintenance of water transport through plant stems. Integrative and Comparative Biology 42: 492–496. Holden, R. 1913. Ray tracheids in the Coniferales. Botanical Gazette 55: 56–65. Höll, W. 1975. Radial transport in rays. In Zimmermann, M. H. & J. A. Milburn (eds.), Transport in plants. I. Phloem transport, 432–450. Springer Verlag, Heidelberg. Johnson, D. M., K. A. McCulloh, D. R. Woodruff & F. C. Meinzer. 2012. Hydraulic safety margins and embolism reversal in stems and leaves: why are conifers and angiosperms so different? Plant Science 195: 48–53. Kedrov, G. B. 2012. Functioning wood (compiled from various articles of G. B. Kedrov by A. C. Timonin). Wulfenia 19: 57–95. Kribs, D. A. 1935. Salient lines of structural specialization in the wood rays of dicotyledons. Botanical Gazette 96: 547–557. Kribs. 1937. Salient lines of structural specialization in the wood parenchyma of dicotyledons. Bulletin of the Torrey Botanical Club 65: 177–186. Lens, F., A. Texier, H. Cochard, J. S. Sperry, S. Jansen & S. Herbette. 2013. Embolism resistance as a key mechanism to understand adaptive plant strategies. Current Opinion in Plant Biology 16: 287–292. Metcalfe, C. R. 1987. Anatomy of the dicotyledons, ed. 2, vol. 3. Magnoliales, Illiciales, and Laurales. The Clarendon Press, Oxford. Metcalfe, C. R. & L. Chalk. 1950. Anatomy of the dicotyledons. The Clarendon Press, Oxford. Metcalfe, C. R. & L. Chalk. 1983. Anatomy of dicotyledons, ed. 2, vol. 2. Wood structure and conclusion of the general introduction. Clarendon Press, Oxford. Meylan, B. A. & B. G. Butterfield. 1978. The structure of New Zealand woods. DSIR Bulletin 122. DSIR, Wellington. Moll, J. W. & H. H. Janssonius. 1909-1936. Mikrographie des Holzes der auf Java vorkommenden Baumarten. 6 vols. Brill, Leiden. Morris, H. & S. Jansen. 2016. Secondary xylem parenchyma—from classical terminology to functional traits. IAWA Journal 37: 1–15. Morris, H., L. Plavcova, P. Cvecko, E. Fichtler, M. A. F. Gillingham, H. I. Martinez-Cabrera, D. J. McGlinn, E. Wheeler, J. Zheng, K. Sieminska, & S. Jansen. 2015. A global analysis of parenchyma tissue fractions in secondary xylem of seed plants. New Phytologist 209: 1553–1565. Morris, H., M. A. F. Gillingham, L. Plavcova, S. M. Gleason, M. E. Olson, D. A. Coomes, E. Fichtler, M. M. Klepsch, H. I. Martinez-Cabrera, D. J. Glinn, E. A. Wheeler, J. Zheng, K. Zieminska & S. Jansen. 2017. Vessel diameter is related to amount and spatial arrangement of axial parenchyma in woody angiosperms. Plant Cell and Environment 1–16. Nardini, A, M. A. Lo Gullo & S. Salleo. 2011. Refilling embolized xylem conduits: is it a matter of phloem unloading? Plant Science 180: 804–811. Ozden, S. & A. R. Ennos. 2014. Understanding the function of rays and wood density on transverse fracture behavior of green wood of three species. Journal of Agricultural Science and Technology B4: 731–743. Panshin, A. & C. De Zeeuw. 1964. Textbook of wood technology. Vol. I. Structure, identification, uses, and properties of the commercial woods of the United States. McGraw Hill, New York. Pfeiffer, H. 1926. Das abnorme Dickenwachstum. Handbuch der Pflanzenanatomie Vol. 9(2). Gebrüder Borntraeger, Berlin. Pittermann, J. and J. Sperry. 2003. Tracheid diameter is the key trait determining the extent of freezing- induced embolism in conifers. Tree Physiology 23: 907–914. Roman-Jordan, E., L. G. Esteban, P. de Palacios & F. G. Fernandez. 2016. Wood anatomy of Cupressus and its relation to geographical distribution. IAWA Journal 37: 48–68. Salleo, S., M. A. Lo Gullo, P. Trifilo & A. Nardini. 2004. New evidence for a role of wood-associated cells and phloem in the rapid xylem refilling of cavitated stems of Laurus nobilis L. Plant, Cell and Environment 27: 341–376. Author's personal copy

Living Cells in Wood 3. Overview; Functional Anatomy of the...

Salleo, S., P. Trifilo, S. Esposito, A Nardini & M. A. Lo Gullo. 2009. Starch to sugar conversion in wood parenchyma of field-growing Laurus nobilis plants: a component of the signal pathway for embolism repair? Functional Plant Biology 36: 815–825. Sauter, J. J. 1966a. Untersuchungen zur Physiologie der Pappelholzstrahlen. I. Jahresperiodischer Verlauf der Stärkespeicherung in Holzstrahlparenchym. Zeitschrift für Pflanzenphysiologie 55: 246–258. Sauter, J. J. 1966b. Untersuchungen zur Physiologie der Pappelholzstrahlen. II. Jahresperiodische Anderungen der Phosphatase Aktivität im Holzparechym und ihre mögliche Bedeutung für den Kohlhydratstoffwechsel und den aktiven Assimilattransport. Zeitschrift für Pflanzenphysiologie 55: 349–382. Sauter, J. J., W. I. Iten & M. H. Zimmermann. 1973. Studies on the release of sugar into the vessels of the sugar maple (Acer saccharum). Canadian Journal of Botany 51: 1–8. Scholz, F. C., S. J. Bucci, G. Goldstein, F. C. Meinzer, A. C. Franco & F. R. Miralles-Wilhelm. 2008. Temporal dynamics of stem expansion and contraction o Scialdone, A. & M. Howard. 2015. How plants manage food reserves at night: quantitative models and open questions. Frontiers in Plant Science 6: 1–6. Scialdone, A., & M. Howard. 2015. How plants manage food reserves at night: quantitative models and open questions. Frontiers in Plant Science 6: 204–213. Secchi, F., C. Pagliarani & M. A. Zwieniecki. 2016. The functional role of xylem parenchyma cells and aquaporins during recovery from severe water stress. Plant Cell and Environment 40: 858–871. Shigo, A. L. 1984. Compartmentalisation: a conceptual framework for understanding how trees grow and defend themselves. Annual Review of Phytopathology 22: 189–214. Trifilo P., P. M. Barbera, F. Raimondo, A. Nardini & M. A. Lo Gullo. 2014. Coping with drought-induced xylem cavitation: coordination of embolism repair and ionic effects in three Mediterranean evergreens. Tree Physiology 34: 109–122. USDA. 1974. Wood Handbook. Government Printing Office, Washington, D. C. Wellwood, R. W. 1962. Tensile testing of small wood samples. Pulp and Paper Magazine of Canada 63(2): T61-t67. Wolkinger, F. 1970a. Morphologie und systematische Verbreitung lebender Holzfasern bei Sträuchern und Bäumen. II. Zur Histologie. Holzforschung 24: 141–151. Wolkinger, F. 1970b. Das Vorkommen lebender Holzfasern in Sträuchern und Bäumen Phyton (Austria) 14: 55–67. Wolkinger. 1971. Morphologie und systematische Verbreitung lebender Holzfasern bei Sträuchern und Bäumen. III. Systematische Verbreitung. Holzforschung 25: 29–30. Zimmermann. M. H. 1971. Dicotyledonous wood structure made apparent by sequential sections: Film E 1975 (film data and summary available as a reprint). Institut Wissenschaftliche Film, Göttingen, Germany.