Printer Friendly

Monocot xylem revisited: new information, new paradigms.

Introduction

Evolutionary concepts in plant anatomy are limited by the fields of knowledge available and taken into account. Certainly we have good descriptive accounts of monocot anatomy in general, based mostly on light microscopy, from the Anatomy of Monocotyledons series begun by C. R. Metcalfe (1960), and now extended by the work of others (e.g., Tomlinson, 1961, 1969, 1983). Cheadle's work on monocots, begun as data summaries (Cheadle, 1942, 1943a, b), was extended, with the collaboration of Hatsume Kosakai (e.g., Cheadle & Kosakai, 1971) to provide family by family examinations of xylem. The end walls and lateral wall pitting of vessels are the focuses of the Cheadle and Kosakai work. Work on monocot xylem has been organized on the basis of systematic groupings, which is ideal for data retrieval (e.g., Wagner, 1977). There is an implication, begun in the nineteenth century by the work of Solereder, that anatomy will yield data useful for the construction of a natural system.

Cheadle (1942) also offers some gradate phylogenetic progressions, based mostly on the end walls of vessel elements: long scalariform perforation plates are considered indicative of "primitive" conditions, simple perforation plates are considered at the opposite extreme, indicators of specialization. The organographic distribution of vessels and their specialization levels were traced by Cheadle and associates. Cheadle's central phylogenetic thesis is that vessels originated in the roots of monocotyledons and advanced upward during evolution into stems, inflorescence axes, and finally leaves (Cheadle, 1942). He also found (1943a, 1943b), not surprisingly, a similar organographic sequence in vessel specialization (many bars to few to none on perforation plates). He envisioned a sort of inexorable trend which could be tracked by means of specialization index numbers. Cheadle's concepts, however, prove to be rather more problematic than has been realized, for reasons that will be presented below.

There are five main sources of new information that now change our ideas on how monocot xylem evolved. The first of these is the construction of molecular trees. Although certainly topologies of these trees are not certain, they have reached sufficient stability and have sufficient levels of likelihood that they must be used as the framework on which we judge ideas of xylem evolution. Prior to Chase et al. (1993), a natural system for the angiosperms was a goal that could only be dimly reached, because anatomy and other indicators do not, as we see in retrospect, form coherent and clearly directional patterns. The ideas of symplesiomorphy, apomorphy, and homoplasy were not features of earlier attempts at a natural system: lists of resemblances were the tool employed, and relationship was judged on the basis on numbers of similarities rather than what character states they represented. The taxonomic groups chosen for comparison sometimes did not even include the groups that now prove, in the light of molecular phylogeny, to be most closely related. In any case, molecular trees now drive the interpretation of xylem evolution, and xylem configuration is no longer a tool in the construction of natural systems, although distinctions of systematic value can still be yielded by xylem.

The second factor that has changed is the widespread use of scanning electron microscopy (SEM). Until recently, use of SEM in studies of monocot xylem was occasional, rather than frequent. SEM proves essential in revealing the occurrence of pit membranes in end walls of tracheary elements, thereby showing that such elements probably should be called tracheids, rather than vessel elements. The production of porose or reticulate pit membranes in these end wall pits, however, has implications not so much for terminology as for the conductive physiology of the xylem. SEM studies, by showing that what hitherto had been regarded as vessel elements are physiologically definable as tracheids invite comparisons with systematics, organography, and ecology, and give us a new understanding of monocot xylem evolution. SEM studies have been changing in methodology (Carlquist & Schneider, 2006), and thickness of pit membranes is now a concern (Jansen et al., 2009).

Earlier students of monocot xylem developed ideas on monocot xylem evolution with little reference to ecology. Xylem is quite often a design for dealing with ecological regimens (Carlquist, 1975). There are multiple plant designs within a given habitat, but each design can be closely cued to xylem function. Ecological information may seem imprecise or highly complex and not capable of analysis by someone interested primarily in xylem, but knowledge of a plant's habitat can point the way to development of focused physiological information. Two families that lie next to each other in phylogenetic trees of monocots, Boryaceae and Asteliaceae, have notably different xylem configurations (Carlquist et al., 2008; Carlquist & Schneider, 2010b). Without knowing that Borya is a "resurrection plant" that grows on briefly moist granite shelves, one would be unable to understand the distinctive vessels and tracheids in its stems. The "primitive" xylem of Astelia, which lacks vessels in stems and leaves but has, in its roots, variously tracheid-like vessels, could not be understood without reference to its highly mesic habitat (often an understory element in cloud forests).

Likewise, habit plays an important role in xylem configuration in monocots. We cannot meaningfully understand why Petrosaviaceae and Triuridaceae lack vessels throughout the plant until we realize that the two families have probably lost vessels independently in response to the heteromycotrophic habit. Lack of vessels in shoots of epiphytic orchids and epiphytic bromeliads relates to the habit, but the differences in the epiphytic habits of the two groups must be taken into account. The succulence of orchids and the tank habit of bromeliads are distinctive adaptations. Vessel diameter in palms relates to whether a species is rhizomatous, erect, or climbing (Klotz, 1977), not to its phylogenetic position within the family Arecaceae.

Ultimately, adaptations in xylem must be determined on the basis of physiological function. Compendia that provide tests of physiological functions (e.g., Zimmermann, 1983; Tyree & Zimmermann, 2002) cannot always be detailed on the anatomy of the plants they study, although plant physiologists are increasingly paying attention to xylem anatomy. Plant physiologists have shown that high root pressure can provide one explanation for the arboreal habit of palms (Davis, 1961) and other monocots (Fisher et al., 1997a, b), and that the valve-like nature of the juncture between stems and adventitious roots in Agave explains how Agave can occupy desert habitats (Ewers et al., 1992). Woody plants are generally easier experimental subjects, so we know much more about the conductive process in non-monocot woody angiosperms than in monocots. Therefore, the discussions of conductive physiology below are less intensive that one could wish. The interesting data that do exist provide motivation for expansion of our knowledge of monocot physiology.

The questions that can now be answered (albeit in a preliminary fashion in some cases) validate the use of a multiple-prong approach to study of monocot xylem. Among these questions discussed in this paper are the following. Were the ancestors to the monocots aquatic? Were the ancestors of the monocots vesselless? What are the advantages and limitations of sympodial stems with adventitious roots, and what role does xylem play in the root/stem juncture? Is monocot xylem constructed for conductive safety or conductive efficiency, or both (and in which species)? What are the advantages of a vesselless stem and leaf xylem, as in so many monocots? What are the special anatomical features of palms and how do they vary with habit and habitat? How do nonpalm arboreal monocots overcome the limitations of lack of a vascular cambium? What are the advantages and disadvantages of the "monocot cambium" and which genera have this kind of lateral meristem? Is progression towards greater vessel "specialization" always progressive, as Cheadle claimed, or can there be reversions, and how can they occur? What restructuring of our ideas on monocot xylem evolution is necessary in the light of molecular phylogenies? What does SEM tell us about vessel elements and tracheids in monocots, and how does that change our concept of what vessel elements are and how they work? What is the syndrome of features associated with the scalariform pitting pattern of tracheary elements? Did vessels originate independently in monocots and dicots? Which basal angiosperms are closest to the ancestral monocots, and what symplesiomorphies might they share? Why do some early-departing clades have more "specialized" xylem than "crown groups"?

Obviously, not all of these questions can be brought to a satisfactory resolution at the present moment. However, original data and synthesis between available information of knowledge from other fields bring us to a new level of conceptual awareness. Although there is much work about monocots recently published, the absence of work on xylem is notable. In fact, data concerning xylem can play a key role in our understanding of the monocots. Although once comparative anatomical data were regarded as elements from which a fallible natural system would slowly be built, the development of molecular systematics has reversed that procedure. DNA-based trees have such high degrees of probability compared to the earlier intuitive natural systems that the newer trees become the framework and testing apparatus for our ideas on how xylem evolves.

Historical Perspective

The first significant contribution to understanding of xylem evolution was that of Bailey and Tupper (1918), who hypothesized a phyletic shortening of fusiform cambial initials. They did not include monocots in their data or in their conclusions, presumably because monocots have no cambium. The Bailey and Tupper (1918) concept was developed in the absence of a reliable phylogenetic tree of the angiosperms, a fact that Bailey (1944) considered a strength of his scheme because it was not dependent on any outside data set. However, Bailey implicitly was aware, by comparing tracheary element length in wood angiosperms with that of gymnosperms and certain fossil groups, that phylogenetic comparison was in some way involved. Lacking any clear phylogenetic tree of woody plants, Bailey and Tupper resorted to a system of inexorable phylogenetic progression stages in xylem, from primitive to specialized.

Bailey soon developed the idea of tracheary element length as a kind of phyletic measuring stick usable for ranking degree of evolutionary departure from primitive character expressions. He soon realized that other characters could be used as phyletic indicators, since he sensed a statistical association among them (Bailey & Tupper, 1918, Table VI). Bailey recognized four categories, based on perforation plate morphology and tracheary element pitting (scalariform perforation plates and bordered pits in tracheary elements defined the ancestral conditions). Bailey handed off the task of elucidating stages in vessel evolution to Frost (1930a, b, 1931). All of these studies were based on non-monocot woody angiosperms. Bailey handed the task of elucidating ray parenchyma and axial parenchyma character state change to Kribs (1935, 1937). Because the symplesiomorphic conditions of all of these were considered statistically linked, Bailey and his co-workers considered that any one of the symplesiomorphic character states (e.g., diffuse axial parenchyma, more numerous bars per perforation plate) could be substituted for vessel element length as a "measuring stick" for phyletic advancement. The quantitative nature of vessel element length as a character made it an appealing first choice as a phyletic indicator, however.

I. W. Bailey's knowledge of non-monocot woods was much more extensive than his understanding of monocot xylem, owing to his forestry background. Bailey handed to Vernon Cheadle the task of determining phyletic trends in monocot xylem. At that time, the data base on monocot xylem was small (e.g., Solereder & Meyer, 1930), probably because the xylem of monocots does not have the commercial importance that angiosperm wood has. Cheadle's five levels of advancement were based mostly on perforation plate morphology. These categories were much like the four categories of Bailey and Tupper (1918, Table VI), differing only in that Bailey and Tupper did not include an all-tracheid condition as Cheadle did. Bailey (1944) was certainly of the opinion that an all-tracheid (homoxylous) wood was ancestral in angiosperms.

Cheadle's groupings, recognized from 1942 onward and even recently (e.g., Thorsch, 2000) were:

0 Tracheids only

1 Vessels with exclusively scalariform perforation plates

2 Vessels with mostly scalariform perforation plates (a few simple plates present)

3 Vessels with scalariform and simple perforation plates about equally common

4 Vessels with mostly simple perforation plates

5 Vessels with exclusively simple perforation plates

Any species or plant portion could be "scored" on this basis, and the scores averaged for families as a whole, so that families could be compared in terms of departure from a hypothetical ancestral condition (0). Protoxylem, early metaxylem, and late metaxylem could also be given separate scores on this basis. One notes that of all the vessel characters given phylogenetic status by Frost (1930a, b, 1931) for woody angiosperms, Cheadle effectively used only perforation plate morphology as the feature in monocots to be considered in detail and consistently in any species or family.

Cheadle's conclusions were stated in a series of principles or dicta, beginning with his first paper comparing a spectrum of monocots (Cheadle, 1942), and stated unchanged years later (Cheadle & Tucker, 1961). These dicta are:

(1) There has been specialization in monocot vessels in the order listed in the above (0-5) scheme.

(2) The organographic specialization of vessels has proceeded progressively from roots (most specialized type of vessels in any given monocot)) to stems, inflorescence axes, and leaves in that order. A few deviations in this sequence (e.g., Dracaena) were noted by Cheadle (1942, 1943a, b).

(3) In any given organ of a vessel-bearing monocot, metaxylem vessels show more "specialization" than those in protoxylem, and late metaxylem is more specialized than early metaxylem.

(4) Longer vessel elements are more "primitive" than shorter vessel elements in monocots. Cheadle stated this principle in his early work (1943a) and retained the idea (Cheadle & Tucker, 1961).

(5) The above trends are held by Cheadle and co-workers to be irreversible.

(6) A monocot group (i.e., family, genus) with more specialized xylem cannot have given rise to a group with less specialized xylem.

(7) Origin of vessels in monocotyledons and in non-monocot angiosperms ("dicotyledons") was independent (Cheadle, 1953).

The above principles are reviewed in the text of the present paper. Cheadle, who died in 1996, did not live to see the enormous impacts that global molecular-based trees of angiosperms (e.g., Chase et al., 1993; Soltis et al., 2000) would have on structural botany. Although Cheadle could not foresee those changes, he did not pursue correlations between structure, physiology, and ecology which were available to him. He and his co-workers worked almost exclusively with light microscopy. Thus, the present account attempts not only to present new and original knowledge of monocot xylem microstructure based on SEM studies, but to synthesize the knowledge gained with light microscopy with information from other fields in an effort to present new ideas of how monocot xylem has evolved. Instead of imposing a scaffold of generalized Baileyan ideas, we must now go in an entirely different direction and use molecular trees as frameworks for organizing xylem patterns, and interpret those patterns in terms of ecology, physiology, and habit.

Materials, Methods, Procedures

The majority of the photographs presented here have not been previously published. Citations in captions document those that have previously appeared in papers. Collection data is cited in captions for figures. Authors of binomials are given in either in captions or in the running text.

The method mostly employed for SEM photographs newly presented here is that described for Orchidaceae (Carlquist & Schneider, 2006). Thick sections of roots, stems, inflorescence axes, and leaves were prepared for SEM study. Thin sections, such as those produced by rotary microtome, prove unsatisfactory because they present limited portions of perforations plates, etc., whereas thick sections can capture the entirety of a perforation plate. In addition, thick sections reveal cell contexts of xylem and the three-dimensional shapes of vessel elements much better than thin sections do. Thick sections also have the advantage of minimizing torsion, which would damage delicate walls, during the handling process. Sections that show end walls of tracheary elements from the inside of an element are more likely to represent unaltered conditions, and are much preferable to sections that show outside of end walls, which have been subject to scraping away of the primary wall by the sectioning process.

Maceration was the technique typically used by Cheadle and associates for study of monocot vessels. It has been used here in a several cases (a few Asteliaceae, Boryaceae, Orchidaceae, and Taccaceae). Maceration is excellent for revealing cell shape and dimensions with light microscopy. Probably Cheadle, having begun with this method, continued it in order to provide data comparable to those he acquired earlier. Macerated cells can also be studied with SEM, and in some families such as Araceae (Carlquist & Schneider, 1998; Schneider & Carlquist, 1998) primary and secondary walls seem relatively unaffected by the maceration process. However, primary walls may experience degradation if the maceration process is prolonged, as is required by material of some other families. Longer maceration times are often necessary with Jeffrey's fluid or any other oxidative reagent capable of dissolving middle lamellae and thereby separating cells. Longer macerating times may be necessary when vascular bundles are sheathed with fibers that separate from each other tardily. Numerous primary wall details have been observed only since Edward L. Schneider and I began using SEM in conjunction with thick sections (Carlquist & Schneider, 2006). Especially important in this regard are the primary walls in pits of end walls of tracheary elements. These are delicate, and can be preserved with reasonable certainty only by means of methods that involve no oxidative or acidic reagents. Materials from macerations are deliberately illustrated here to contrast their probable degrees of loss of pit membrane portions with the appearances obtained from sections of alcohol-fixed material.

Materials to be studied were fixed in 50% aqueous ethanol. Sections (about 1-2 mm in thickness) were cut manually with a single-edged razor blade. Sections were then subjected to three changes of distilled water at 50 [degrees]C in order to remove extraneous substances. Sections were then placed between pairs of glass slides and pressure applied with a clip in order to assure flatness of the dried section. Drying was accomplished by placing the glass slides so assembled on a warming table at 50 [degrees]C until drying has occurred. Dried sections were then examined according to the usual techniques.

The thick sections offer the advantage not merely of revealing large portions of or the entirety of a perforation plate, but in being able to reveal it as seen from the inside of a cell, or in the form of oblique sections. These views reveal with greater certainty the presence of pit membranes in end walls of tracheary element. Some perforation plates may be separated by the sectioning process into halves representing the two component cells, but such split or scraped perforation plates reveal portions of the primary wall to various degrees, and are. Sections that reveal intact perforations (as opposed to split) perforation plates are the most valuable because they show, relatively free from artifacts, what the conductive stream in xylem encounters when traversing an end wall. Intact perforation plates are visible, from the inside of a vessel element, and thus a longisection that exposes the lumen is required.

New Data, New Contexts

What is a Vessel?

Anatomical Considerations

Vessel elements are defined on the basis of presence of perforation plates on end walls, composed of one or more perforations. This classical definition is, ironically, a statement of what is not present: primary walls are absent in the perforations. But is our knowledge of this reliable? The definition is based on light microscopy, which in the case of macerated cells (or perforations plates as seen in face view in sections) rarely can show presence of pit membranes in the perforations, even with particular staining methods. The inference of absence of pit membranes in these cases has been made on the basis that perforations are larger than the pits on lateral walls of a particular vessel element (Fig. la, b; Fig. 4f). If the perforations are larger, and a perforation plate can thereby be declared to be present, the absence of pit membranes can be inferred with reasonable certainty (but pit membrane remnants may be present in some scalariform perforation plates: viz, Carlquist, 1992a). The oversimplified drawings of Kosakai (e.g., Cheadle & Kosakai, 1971) tend to suggest that perforation plates can be delimited clearly. When long scalariform perforation plates are examined with SEM, however, the distinctions between end walls and lateral walls may prove elusive.

Interestingly, in one supposed vessel element (Anigozanthos rufus Labill., Haemodoraceae), Cheadle (1968) superimposed a pattern on the perforation plate, indicating his uncertainty that pit membranes actually were lacking. This indication is, however, a unique instance. Uncertainty about whether a tracheary element is actually a vessel element or a tracheid led Fahn (1954), who used light microscopy, and Wagner (1977) to employ the term "vessel tracheid" to call attention to apparent intermediacy. This usage calls attention to the problem, but does not add new information. The term is also misleading, because vessel element, not vessel, would be the counterpart to tracheid. The nature of instances of apparent intermediacy could not be elucidated in the era when light microscopy was the sole tool for viewing xylem. Pit membranes in Astelia (Fig. 1c-g) demonstrated with SEM cannot be seen with light microscopy. Transmission electron microscopy is also a valid tool for revealing presence of pit membranes in tracheary elements, but it has been used only rarely in monocots (e.g., Thorsch, 2000), probably because lack of commercial importance of monocot xylem. SEM has also been relatively little used to date in the study of monocot xylem.

Cheadle (1942) referred to passage of India ink particles through perforation plates of monocots as a criterion for discrimination between vessel elements and tracheids. By doing this, Cheadle implicitly recognized that pit membranes in some form might be present in "perforations." India ink particles are about 1 [micro]m in diameter. By extension, one could also use colloidal latex microspheres, which are now commercially available in uniform-diameter populations. Such microspheres would best be detected by study of vessels with SEM after the latex had been taken up by a plant or injected--a complicated and prolonged procedure. We now know that size of porosities with end-wall pit membranes varies considerably (e.g., Fig. 3). What is the porosity size that relates to a physiological distinction? Passage or non-passage of air bubbles within a transpiration stream might represent an example of such a distinction.

Fahn (1954) rejected the passage of India ink particles through a perforation plate as an aid in deciding whether a cell is a vessel element or a tracheid. Instead, he entertained the idea that by pressing a needle on a cover slip of a maceration preparation, one can perform a test. If this action displaces some of the bars, he believes that no pit membranes interconnect the bars; if the action results in equal spacing of the bars, he thought that lack of pit membranes was indicated. This test may have been appropriate in its time, but it has been supplanted by SEM studies.

Some large monocot families have long scalariform end walls on tracheary elements (Bromeliaceae, Orchidaceae, and Pandanaceae, for example). In these families, SEM is the only reliable method to decide both whether pit membranes are present in end wall pits and what the membrane microstructure is (porous, reticulate, etc.). Thus, the only secure method for deciding whether tracheary elements or vessel elements are present in particular species becomes an elaborate procedure, available to few. This will appeal to some workers as an untenable situation in terms of terminology, but development of mutually exclusive terms is not a realistic or even desirable goal in this instance. Instead, demonstration of the structural continuum and thereby the evolutionary and physiological status of tracheary elements in monocots becomes a much more important goal. Fahn (1954) offers the term "vessel tracheid" for intermediate tracheary elements. Such a proposal may well be a better choice that attempting to reinstate definitions that had their origin in light microscopy and are not applicable when microstructure is taken into account.

Astelia (Asteliaceae, formerly a subfamily of Liliaceae) has been regarded as having vessels only in roots (Cheadle & Kosakai, 1971). On the basis of Fig. la and b, one could judge vessel elements to indeed be present. However, these two figures are based upon macerations, and delicate primary wall material might have been removed from the perforation plates by that process. One can find, within a single root, perforation plates that vary in terms of presence and size of porosities in the pit membranes or pit membrane remnants of perforations (Fig. 1a-e). In Fig. 1d and e, the end wall membranes are entirely intact. Possibly these could be immature, but they were not taken from apical portions of the roots, and such plates may be found here and there in Astelia root xylem. Figure 1a-e all show narrow bars characteristic of perforation plates.

End walls in stem tracheary elements of Astelia show little difference between end walls and lateral walls in the secondary wall patterns that delimit pits. There are, however, threadlike remnants of primary wall material to a greater (Fig. 1f) or lesser (Fig. 1g) extent.

Can pit membranes be either present or absent within a given root portion of Astelia? Presumably so, although some caution should be observed. The perforations of Fig. 1a-b are from macerations, and the oxidative properties of macerating fluid can result in loss of primary wall material not only from end walls, but from lateral wall pits as well (see Fig. 5a-d). This happens when xylem is enclosed within fibers, because the prolonged treatment needed to macerate the fibers also can damage primary walls of xylem elements.

The possibility remains that tracheid-like vessels can be intermixed with vessels with perforations lacking pit membranes. SEM study of sectioned material is ideal at revealing pit membranes in perforation plates, but it can sample only a small number of perforation plates, leaving uncertain what a large population of vessel elements might show. Sections in which one can view perforation plates from the inside of a tracheary element (Fig. 1 c) are ideal. Of equal value are oblique sections of end walls (Fig. 1f: compare the threads of the perforation plate at right to the solid sheets of wall material to the left).

Frequently in sectioning, tracheary elements are split apart, rather than sectioned down through a lumen. This may result in "scrape away" effects, in which various amounts of primary wall material are removed, depending on how deeply the blade edge cuts into any given pit membrane. Cellulosic fibrils are often revealed well (Fig. 1g), but one is uncertain how much wall material has been removed. For accurate information about intact pit membranes, views from the lumen sides of tracheary elements or oblique sections of end walls are much more reliable.

Given these considerations, natural patterns of pit membrane retention can differ from one monocot to another. The inflorescence axis of Canna (Fig. 2a-e) represents conditions somewhat different from those of Astelia. The perforation plate of Fig. 2a is viewed from the inside of the tracheary element, and has not been affected by the sectioning process. Rather coarse microfibrillar pit membrane remnants are present. In Fig. 2b, the wall between two adjacent vessel elements has been split, indicating that the fibrils belong not just to one of the two adjacent tracheary elements, but to both, as one would expect. Figure 2c-e are sections that show most of the fibrils in a perforation (or pit) of end walls. The fibrils are intercontinuous from one perforation to another (Fig. Id). Larger holes in the network (Fig. 1c, e) may be indicative of removal of a few strands by sectioning.

In the root perforation plate of Canna, no primary wall material is present in most of the perforations (Fig. 2f, far right), although pits transitional between lateral wall pitting and perforation plates (Fig. 2f, bottom center) may retain pit membranes (which are fractured in this micrograph, the fracturing presumably an artifact). Scalariform perforation plates are not perfectly delimited from lateral wall pitting, although drawings of perforation plates often suggest that they are.

Orchidaceae offer additional examples of transitions between vessel elements and tracheids. In Phalaenopsis (Fig. 2a-d), one can se a progressively diminished degree of poroussness of end wall pit membranes, beginning with roots (Fig. 3a), then stems (Fig. 3b), and finally inflorescence axes (Fig. 3c). Lateral walls of vessels (or vessel-like tracheids) in Phalaenopsis (Fig. 3d) show no pores in the pit membranes. Roots of Vanilla have only small pores in pit membranes (Fig. 3e: the fracturing should be disregarded). Vessels illustrated for Stenoglossa (Fig. 3f) are from sectioned material, and this appearance, common in sectioned material, must be considered in the light of the "scrape away" effect (for an example of this in woody material, see Jansen et al., 2009, Fig. 2).

Epidendrum roots (Fig. 4a-c) have abundant longitudinally-oriented strands in perforations. Stems from the same Epidendrum plant (Fig. 4d-e) have more nearly intact but prose membranes. In Odontoglossum (Fig. 4f), one can see a "classical" perforation plate, although some pit membrane remnants might have been removed by the macerating process. Sections of Odontoglossum stems (Fig. 4g) have pores in pit membranes only where pit membranes are scraped away by the sectioning process; the pit membranes are otherwise intact.

By contrast, roots of Orcbidaceae subfamily Apostasioideae (sometimes cited as Apostasiaceae) can have circular perforation plates that do not occupy the entirety of an end wall (Fig. 5b c). Tracheids may also be encountered (Fig. 5a). There is little difference between roots and stems with respect to tracheary end wall structure. One must remember, however, as noted by Wagner (1977), that little of the family has been sampled.

[FIGURE 4 OMITIR]

Ophiopogon (Asparagaceae, subfamily Dracaenoideae) roots sometimes have only a small number of pores in tracheary element end wall pit membranes (Fig. 5e, f). This is true whether the end wall has experienced some sectioning (Fig. 5e) or is intact (Fig. 5f).

Narrow tracheary elements of Typha (Typhaceae) roots should be called tracheids because they retain a primary wall meshwork in the end wall pits (Fig. 6a-b, e-f). Wider tracheary elements of Typha roots (Fig. 6c-d) can be called vessel elements because the end walls lack pit membranes in most perforations. At the end of the perforation plate (Fig. 6c-d) where there is a transition to lateral wall pitting, pit membrane remnants are present to various degrees, however. In stems of Typha, end walls of tracheary elements have reticulate pit membranes (Fig. 6e-f). This contradicts the data of Cheadle (1942) and Wagner (1977) to the effect that vessels occur throughout the plant in Typhaceae.

Similar considerations apply to Cyclanthaceae (Fig. 7a-c) and Pandanaceae (Fig. 7d-i), families that apparently are closely related to each other (Chase, 2004) In Carludovica of the Cyclanthaceae, roots have long vessels with many-barred scalariform perforation plates (Fig. 7a, b). In these, perforations can be said to be present on the basis that most of their area is devoid of pit membranes, but pit membrane remnants can be found at the lateral ends of perforation plates (Fig. 7c). Wagner (1977) records vessel presence in roots of Cyclanthaceae, but only tracheids in stems. That view can be maintained on the basis of the present study (if one takes the viewpoint that a reticulate pit membrane characterizes a vessel element rather than a tracheid).

The pattern seen in Typha is present in Freycinetia of the Pandanaceae also. Roots of Freycinetia have long scalariform perforation plates (Fig. 7d) with some pit membrane remnants at lateral ends of perforations (e.g., Fig. 7e, extreme left). In stems of the same Freycinetia, there are clearly porose pit membranes in end walls of tracheary elements, whether seen where an adjacent vessel has been sectioned away (Fig. 7f) or whether one is seeing an intact pit membrane from the inside of a tracheary element (Fig. 7g). This is also true of a stem of Pandanus studied (Fig. 7h-i). Cheadle (1942) and Wagner (1977) claim presence of vessels throughout the plant in Pandanaceae, but that is not confirmed on the basis of the present SEM studies.

Lapageria (Philesiaceae) adds further dimensions to this pattern. SEM micrographs reveal that in the roots of Lapageria, pit membranes are present in presumptive end walls of tracheary elements, whether viewed from inside of the element or from the outside surfaces in which adjacent tracheary elements have been sectioned away (Fig. 8a-b). The pattern of the pit membranes in Fig. 8a-b) suggests a dense reticulum, with some pores present. The number of bars on the end walls of Lapageria root tracheary elements is quite high (150-700 according to Fahn, 1954, who calls the elements "vessel tracheids"). Loss of pit membranes in such narrow pits of an end wall is highly improbable, based on SEM studies of such long plates. Consequently, end walls of Lapageria are virtually indistinguishable from lateral walls in terms of pit morphology. The criterion for identifying end walls used in the SEM studies reported here is that end walls traverse an element diagonally as seen in a longitudinal section. On the basis of the present study, Lapageria can be called a non-aquatic vesselless angiosperm (it grows in moist forests, but in soil that is never inundated).

Roots of Taccaceae have very long perforation plates. Fahn (1954) reports 30-200 bars on end walls of "vessel tracheids" of roots of Schizocapsa plantaginea Hance, 80-100 on those of Tacca paxiana W. Limpricht, and 90-300 on those of T. palmata Blume. I have found about 40-100 bars on root tracheary element end walls of T. leontopetaloides (Fig. 8c) and T. chantrieri Andre, and at least 200 on those of T. integrifolia W. M. Curtis (original data). Such a large number of bars, combined with such narrow perforations, is incompatible with complete hydrolysis of pit membranes from the end wall pits, so various kinds of pit membrane remnants occur. Material of T. leontopetaloides from macerations (Fig. 8c) does not show such remnants: a result of the chemical removal of primary wall material by the acidic and oxidative qualities of the macerative fluid. However, thick sections of liquid-preserved material of T. chantrieri, T. integrifolia, and T. leontopetaloides show a range of appearances, from intact but somewhat porose membranes (Fig. 8d, e) to pit membranes with both large and small pores (Fig. 8f, g). all of the micrographs of T. leontopetaloides are from portions of a single plant. Even in older root portions, pit membrane remnants are retained. No end walls of sectioned roots lack pit membranes or pit membrane remnants of some kind, so on a functional as well as a morphological basis, the tracheary elements of roots are closer to the tracheid end of the gamut than to the vessel element end.

The lateral walls of vessel elements in T. integrifolia roots are notable for consisting of scattered circular bordered pits. These connect to fibriform tracheids that bear circular pits with prominent borders.

The difference between vessel end walls and tracheid end walls is not very great in a number of monocot species. It is merely a matter of persistence of pit membranes that are sometimes so networklike (as in Typha) that the holes in the meshes occupy much more area than do the strands of the network. In monocots, scalariform pitting on end walls of tracheids closely resembles the scalariform perforation plates on end walls of "less specialized" vessel elements. Varied thicknesses of pit membranes in vessel element have been reported by Jansen et al. (2009), who find that larger pores may be found in thinner pit membranes.

If one envisions a "primitive" vessel in developmental terms, a nonporose pit membrane (but traversed by micropores of the plasmodesmata) is present at maturity of the cell. As the protoplast vanishes, dissolution of soluble (presumably mostly pectic) portions of the end wall occurs. This leaves a remnant reticulum of cellulosic fibrils in some end wall pits. These reticula can be swept away by the conductive stream to various extents. The wider the tracheary element, the lower the likelihood that the pit membrane reticulum will remain intact, because stresses on an elongate pit membrane are greater than those on a short pit membrane. The wider the tracheary element, the less conductive resistance it has and the greater the likelihood that the pit membranes in its end wall pits will be swept away, producing a vessel element, by a form of hydrolysis (Butterfield & Meylan, 1982).

Physiological Considerations

Evidence from comparative anatomy has justifiably been taken as indicating that vessels confer an advantage to conduction. However, Hacke et al. (2007) and Sperry et al. (2007) issue a caution, claiming that removal of pit membranes from an end wall does not by itself confer much conductive advantage, although vessel widening and simplification of the perforation can be still be considered appreciably advantageous. Sperry et al. (2007) say, "primitive scalariform plates were major obstructions to flow, accounting for 50 % of the total flow resistivity on average." Ellerby and Ennos (1998), on the other hand, reported that vessel element end walls, whether scalariform or simple, confer a small portion of resistivity to conduction (0.6-18.6 %) when compared to the resistivity caused by the lateral walls. However, what has not and cannot be measured is the conductive capabilities of a vessel element and an equivalent transectional area of tracheids in a given species. If such a measurement were possible, it would undoubtedly show that the vessel holds an advantage over an equivalent transectional area of tracheids. Long scalariform perforation plates do provide resistance to flow, but they may have the advantage of decreasing likelihood of air embolism formation and of promoting recovery from embolisms and aiding refilling, based on the ideas of Kohonen and Helland (2009). Ellerby and Ennos (1998) indicate that perforation plates do not confer nearly as much resistance (0.68.6 % of the resistance offered by the vessel), and are much less important in this respect than the vessel walls. Widening of the lumen confers a major advantage (the fourth power of the diameter increase) to vessels (Tyree & Zimmermann, 2002), a fact that is easily reflected in the wide diameter of earlywood vessels.

The function of perforation plates with a limited number of bars, which are found in a number of palms (Klotz, 1977), is not clear. They may be sites for resistance to high positive or negative pressures in vessels (Carlquist, 1975) or may even confer mechanical strength of some other sort.

Wider vessels with simple perforation plates offer potentially increased vulnerability to the conductive system. Air bubbles involved in embolisms can spread from one vessel element to another. Perforation plates, even simple ones, may tend to restrict air bubbles, as compared to an ideal continuous smooth capillary (Slatyer, 1976; Sperry, 1985, 1986; Ewers, 1985; Kohonen & Helland, 2009). Scalariform perforation plates would be expected in this scenario to confine embolisms to individual vessel elements, especially if they have pit membranes in the end walls.

Removal of air embolisms and mechanisms for recovery of the water columns in vessels have received considerable attention in recent years (Clearwater & Goldstein, 2005; Pickard & Melcher, 2005; Holbrook & Zwieniecki, 1999), and clearly is widely operative. Root pressure is pronounced in some monocots (Davis, 1961) and is a widespread phenomenon in monocots as well as certain non-monocots (Ewers et al., 1997; Fisher et al. 1997a, b). Most monocots are within the height range where root pressure would be effective. Information on refilling of cavitated vessels in grasses is offered by McCully et al., (1998) and Stiller et al. (2005).

Hacke et al. (2007) refer to "cryptic vessels," which have greater porousness of end walls than typical tracheids, but do not clarify this concept. Feild et al. (2000) figure tracheids with pit membranes lacking in Amborella, but these are artifacts, because intact porose pits in end walls of Amborella can be found (Carlquist & Schneider, 2001; Hacke et al., 2007). The pit membranes of Amborella are very delicate and break easily. This is also true in Bubbia, a genus of the vesselless Winteraceae (Carlquist, 1983). Pit membrane thickness may be important in study of conduction of vessels, but data from monocots is lacking. Our ideas about the physiology of conduction are based largely on woody angiosperms and conifers. Presumably these concepts can also be demonstrated with monocots (e.g., Sperry, 1985, 1986), but all monocots are not alike in xylem anatomical formulae or in quantitative characteristics (e.g., Fisher et al., 1997a, b).

A null hypothesis (that scalariform perforation plates have no function, but are a feature that has persisted from ancestral species) does not seem likely, because structural evolution is too efficient for mass persistence of a functionless character state. No functionality, however, is implied by the work of Cheadle (1942 et seq.), who presents vessel evolution in monocots as an inexorable process of perforation plate simplification. Structures such as scalariform perforation plates, which represent considerable expenditure of photosynthates, are not likely to be present for relictual reasons.

Vessellessness in Monocots: A Pervasive and Important Theme

How common is vessellessness in monocots, systematically and organographically? What relationships exist between vessellessness and ecology and conductive physiology? What relationships exist between characteristics of plants and organs (e.g., succulence) an vessellessness? Are there monocots in which primary xylem and early metaxylem are vesselless but metaxylem contains vessels? Cheadle (1942, 1943a, b) stressed evolutionary development of vessels within monocots systematically and organographically. However, should one not stress, instead, the inverse: the retention of all-tracheid systems in many monocot organs, and try to establish why this retention has occurred? Vesselless woody angiosperms are few and geographically restricted, but if one looks at a landscape and knows which monocots lack vessels one may be surprised at how much vesselless vegetation one is seeing. This is true even in a grocery store: the edible parts of onions, garlic, and asparagus, for example.

Vessellessness is very common at the organ level in monocots. There are many monocot species in which all tracheary elements of stems and leaves have pit membranes on all pits. Such elements fit the traditional definition of a tracheid, but that is a problem if SEM is required to establish pit membrane presence in end walls. Any other defmition (e.g., degree of porousness of the pit membrane) is equally problematic, also relying on SEM data that would take decades to accumulate. The traditional light microscope definition--a scalariform end wall in which the perforations are wider (and the bars between them are narrower) than the pits of lateral walls of tracheary elements-- will probably continue to be used because it can be applied so easily in light microscopy and can be demonstrated in many species. Examples of intermediacy will, in this case, not be stressed. SEM data now available suggest that in most cases in which there is little difference between end walls and lateral walls of tracheary elements, pit membranes are likely to be present on end walls. If this is true, vessellessness is much more common in monocot stems and leaves than the listings of Cheadle (1942 et seq.) and Wagner (1977) would lead one to believe.

Examples are presented above for Asteliaceae (Fig. 1), Cannaceae (Fig. 2), Orchidaceae (Figs. 3, 4), dracaenoid Asparagaceae (Fig. 5e-f), Pandanaceae (Fig. 7d-f), Philesiaceae (Fig. 8a-b) and Taccaceae (Fig. 8c-g) These represent just a small sampling of species I regard as likely to have pit membranes in end wall pits rather than having perforations, as reported (Wagner, 1977). To be sure, there is progressively less porousness within an end wall pit membrane of a single plant as one goes from root to inflorescence axis in Epidendrum and Phalaenopsis, suggesting a decrease in conductive ability as one goes from root to inflorescence axis, but in the form of a minor gradation of tracheid microstructure. To be sure, the root of Odontoglossum (Fig. 3f) seems to have perforation plates, but the illustration is from a maceration, a technique that could remove pit membrane remnants.

The pattern of vessel presence in roots combined with tracheids elsewhere in the plant is common in monocots, and one needs to account for the significance of this pattern. No monocots are reported by Wagner (1977) to lack vessels in roots except for a few families in ecologically special circumstances. These include families of submersed aquatics (e.g., Aponogetonaceae, Ruppiaceae, Zosteraceae). Also notable in this regard are families which are heteromycotrophic, such as Petrosaviaceae, Triuridaceae, and achlorophyllous Burmanniaceae and Orchidaceae (Carlquist, 1975; Wagner, 1977). These families are denoted by small circles to the left of family names in Fig. 15.

The species studied here present some interesting examples. Typha does have wider metaxylem vessels that lack pit membranes in roots but also narrower metaxylem vessels in which end wall pits still have pit membranes (Fig. 6). Similar situations occur in Cyclanthaceae and Pandanaceae (Fig. 7) when microstructure is studied. This suggests a relationship between tracheary element diameter and degree of end wall primary wall retention. In woody dicots, vessels tend to be wider in roots than in stems (Patel, 1965), so a similar trend is not unexpected in monocots. Thus widening of tracheary elements could account, in part, for presence of vessels in roots of monocots which lack vessels in stems and leaves. It might also account for the apparent lack of vessels in the roots (Fig. 8a-d) of Lapageria (Philesiaceae), and the lack of vessels in stems of the palm Phytelephas (Klotz, 1977), and the apparent lack of vessels throughout the plant in most Taceaceae (see original data above).

We are still confronted by the question of why tracheids or tracheid-like vessels should occur in these plants. There are two prime questions: can there be secondary vessellessness within particular organs? And what ecological circumstances favor the presence of vesselless conditions in particular organs or the occurrence of vessels in others? These questions have been sidelined in earlier studies. Cheadle (1942 et seq.) merely believed in progressive inexorable acquisition of vessels, and thought that acquisition of vessels was an irreversible trend. Cheadle also averred that links between xylem and ecology and between xylem and habit in monocots should be sought "after the data on xylem in monocots is in" (V. I. Cheadle, personal communication, 1994), and regarded attempts to find such correlations (Carlquist, 1975) as premature.

The heteromycotrophic monocots use a network of fungi instead of root hairs as a method of absorbing water, and all grow in notably mesic forest floor locations rich in leaf litter and have limited aboveground stature. Under these circumstances, tracheids suffice quite well for conduction and, in fact, in all of these taxa, the tracheids are narrow (Carlquist, 1975). The systematic distribution of the heteromyeotrophic monocots (Fig. 14) suggests that there has been vessel loss and that autotrophic ancestors probably had vessels in roots.

In some other monocots mentioned above in which tracheids are more pervasive, and vessels less common than had been thought by Cheadle (1942) and Wagner (1977), highly mesic habitats are a common denominator. This is true in Asteliaceae, Cyclanthaceae, many Pandanacaeae, and terrestrial Orchidaceae, for example.

Aquatic habitats are, of course, the ultimate in making minimal transpiration demands on a plant. Therefore, the many families and genera that are vesselless among aquatics, are, not surprisingly, vesselless. These occur in Alismatales (or Alismatidae) mostly. For these plants, vessels have little or no selective value except in relation to fluctuating levels of moisture, as is the case in pond or stream margins (e.g., Sagittaria), where brief periods of lowered water availability may correlate with the increased conductive rates, which may rapidly reverse any embolisms that form on hot, dry days. Although Alismatales are an early branch in the monocot tree (Fig. 14), they may not be ancestrally vesselless. Some vesselless monocots, such as Zosteraceae, may solve the low-oxygen content problem of an underwater habitat by living in areas subject to wave action, and thereby maximal water oxygenation. Other submersed aquatics have solved the problem of low oxygen availability in standing water by developing air circulation patterns within the plant body, as is known for non-monocots such as Nymphaeaceae and Menyanthaceae. The complexity of these specializations correlates with the rather small number of families and genera that have adapted to the submersed aquatic habitat. We should consider the possibility that the xylem and the air circulation systems of these plants represent some apomorphic features, and are not wholly symplesiomorphic, even though they may retain antique DNA sequence patterns.

Succulence and lowering of transpiration by thick cuticles, sunken stomata, drought deciduousness of aerial portions of leaves (e.g., Allieae) and C4 photosynthesis (e.g., Silvera et al., 2010) are mechanisms that should be studied in conjunction with vessellessness of particular organs in monocots. "Protection from transpiration" (= high diffusive resistance of leaf surfaces; condensed leaf forms) and "internal mesomorphy" (= succulence) are themes that should be studied in relation to vessellessness in stems and leaves of monocots.

Tracheids Coexisting with Vessel Elements

Tracheids that form the background of vessel bearing woods (e.g., Cornaceae, Hamamelidaceae, many Rosaceae) are a commonly encountered phenomenon in the woods of many woody angiosperms. Xylem in which tracheids as well as vessel elements occur alongside each other in a single vascular bundle occurs in monocots, but has not been sufficiently appreciated. Certainly co-occurrence of tracheids and vessels together has been reported (Cheadle, 1942; Fahn, 1954; Klotz, 1977; Wagner, 1977). When both are present together, a kind of intergradation between the two cell types may be characteristic.

A strong division of labor between co-occurring tracheids and vessel elements is present, however in Borya (Carlquist et al., 2008). The stems of Borya have scalariform perforation plates (Fig. 9a-c). The bars are thin to extremely tenuous, and often collapse in cell macerations. In vessels, the perforation plates are well differentiated from lateral wall pitting, which consists of alternate circular pits (Fig. 9b, upper right).

Most xylem cells are narrow tracheids with one to three rows of prominently bordered alternate circular pits (Fig. 9d-f). The tracheids are fusiform, in contrast to the vessel elements. In fact, the xylem of Borya stems in a maceration looks much like the xylem of a woody angiosperm.

To understand the co-occurrence of two such contrasted cell types in stems, one must know that Borya is an Australian "resurrection plant" that grows on granite shelves which may be wet and dripping during rains, but which are dry for most of the year. The vessel elements offer the potential of rapid supply of water to the foliage with the initiation of the rainy season. The tracheids are thick-walled, and can probably maintain water columns even under water tension during the dry season.

The xylem of Borya is not what one would expect from an early-departing, near-basal branch of Asparagales if one thinks in terms of gradual phylogenetic progressions as Cheadle did. Instead, the xylem design shows radical design suited for a special ecological situation.

The lateral walls of wide, vessel-like tracheary elements (which possess pit membranes or pit membrane remnants) in roots of Tacca integrifolia have scattered circular lateral wall pits. These circular pits connect to fibriform tracheids with circular bordered pits (unpublished data). Thus, more than one kind of functionally imperforate tracheary element can co-occur in T. integrifolia roots. Tacca integrifolia is an understory plant of moist tropical forests, and clearly unlike Borya. The root xylem of Tacca integrifolia may have counterparts in other wet forest monocots, such as Lapageria. Our knowledge of such monocots using SEM is as yet rudimentary.

Arecaceae is an interesting family with respect to co-occurrence of vessels and tracheids. Klotz (1977) indicates imperforate tracheary elements (= tracheids) present in early metaxylem of roots, stems, and leaves of all of the palms he studied. In most species, late metaxylem in these species has vessels. This is an interesting kind of co-occurrence that has, like the quite different xylems of Borya and Dracaena, implications for retaining conductive safety (tracheids resist spread of embolisms) with conductive efficiency (the metaxylem vessels of palms are few per bundle and notably wide).

Lateral Meristem Activity in Monocots and Its Implications

Lateral meristems in monocots (called "monocot cambia" here so as not to be confused with types of cambial activity in non-monocots) have been studied by various workers, notably Cheadle (1937), Tomlinson and Zimmermann (1969), and Rudall (1991). The genera in which monocot cambia have been recorded include the following (taxonomy according to APG III, 2009, and the tree of Fig. 15). The term "monocot cambium" is equivalent to "secondary thickening meristem" as used by Rudall (1991). Rudall is doubtful that the monocot cambium is equivalent to secondary thickening meristematic activity in non-monocot angiosperms and Gnetales, and indeed, it is not. The process by which a "master cambium" arises and gives rise to conjunctive tissue and to vascular cambia, which in turn, produce xylem and phloem, is quite a different process (Carlquist, 2007), and thus the contrasting terms "monocot cambium" and "master cambium: are used here. Monocots either have no cambium or a cambium-like layer in bundles that, in fact, is permanently dormant and produces no vascular tissue (Carlquist, 2007). According to the most recent compilation of Rudall (1995), monocot cambia are found in:

Asparagales
  Iridaceae
    Aristeoideae
      Aristea
    Nivenioideae
      Klattia
      Nivenia
      Schizostylis
      Witsenia
  Xanthorrhoeaceae
    Xanthorrhoeoideae
      Xanthorrhoea
    Asphodeloideae
      Aloe
      Gasteria
      Haworthia
      Trachyandra
  Asparagaceae
    Aphyllanthoideae
      Aphyllanthes
    Agavoideae
      Agave
      Beaucarnea
      Calibanus
      Chlorophytum
      Dasylirion
      Dracaena
      Furcraea
      Hesperaloe
      Hesperoyucca
      Nolina
      Pleomele
      Thysanotus
      Yucca
    Lomandroideae
      Cordyline
      Lomandra


The above listing is not as simple as it may seem. Monocot cambia produce relatively few if any secondary bundles in the genera of Dasypogonaceae; Baxteria, Calectasia, and Kingia may not belong on this list and are not so included in the survey of Rudall (1995). Formation of tracheids, either by differentiation of pre-existing parenchyma cells, or preceded by a few divisions that could be considered rudimentary meristematic activity, has been reported in root-stem junctions in some Bromeliaceae, Commelinaceae, and Zingiberales, for example (Tomlinson & Zimmermann, 1969; Rudall, 1991). These latter instances need further study.

The admirable essays by Tomlinson and Zimmermann (1969) and Rudall (1991) make the point that monocot cambium is a continuation of the primary thickening meristematic activity which enlarges the meristematic zone at the shoot tip. The primary bundles may still be maturing at the same level where lateral meristematic of the monocot cambium is already in progress; however, there may be a discontinuity between the two processes. Stevenson (1980) shows that the two processes can be intercontinuous in seedlings of Beaucarnea, but discontinuous in the adult plant. Tomlinson and Zimmermann (1969) make the interesting, if minor, point that addition of more numerous bundles on the lower surface than on the upper surface of a slanting stem serves the purpose of reaction wood. They report that the monocot cambium can originate both inside and outside of the endodermis in roots of Dracaena, even within a single section.

The ontogeny and mature stems of Yucca brevifolia exemplify secondary bundle formation (Fig. 10). A meristematic layer forms in the cortex of a stem (Fig. 10a, pointers). Products of this meristem are radially aligned, and therefore can easily be distinguished from the primary cortex (Fig. 10a, right) and the primary part of the stem internal to the cortex. Primary cortex cells are not radially aligned. In younger stems of Yucca brevifolia, periderm develops from periclinal divisions in the outer primary cortex. As the stem increases in size, the periderm and primary cortex become broken into functionless segments but are retained on the stem. As this happens, new periderms are initiated within secondary cortex.

The monocot cambium produces radial files of meristematic cells internally (Fig. 10a-b). Vascular bundles are initiated (Fig. 10b, vbi) by means of divisions within these radial files. Two early stages are indicated in Fig. 10b, and one of these is shown enlarged in Fig. 10c. Divisions continue (left half of Fig. 10b) until an optimal strand of procambium-like cells is achieved (Fig. 10a, left). These then differentiate into collateral bundles, with phloem external (Fig. 10d, p). The xylem (Fig. 10d, x) part of each bundle is much larger than the phloem and consists wholly of tracheids.

The Yucca brevifolia pattern, with variations, occurs in other monocots with secondary growth. In Dracaena deremensis (Fig. 11), an early stage in secondary activity is depicted. The monocot cambium has produced only about two layers of secondary cortex at this stage. Toward the inside, a single series of secondary bundles has been produced. These bundles are amphivasal rather than collateral. Phloem (p) and a tracheid (t) are shown for one of the secondary bundles in Fig. 11a. Only a few tracheids are mature in these secondary bundles, so that the amphivasal nature is not conspicuous. The primary bundles (Fig. 1 a, left half) are collateral, with phloem (p) external) to two or more tracheids (t). In addition, the external face of the primary bundles consists of extraxylary fibers (f).

The features of Dracaena stem bundles are illustrated more conspicuously by the bundles of Cordyline (Fig. 11b-c). Primary bundles (Fig. 11b) tend to be collateral, with protoxylem (px) internal to the central phloem strand, but metaxylem (mx) external to it. All xylem cells are tracheids. The secondary bundles (Fig. 11c) are clearly amphivasal, with tracheids surrounding a central strand of phloem. Note that because secondary bundles are derived from meristematic (procambium-like) cells that do not elongate, as do those in primary stems, the tracheids can all be considered to resemble metaxylem tracheids, and the elements formed in secondary bundles are pitted rather than with annular or helical thickenings.

In the dracaenoid genera, Cheadle (1942) reported vessels only in the roots, with an all-tracheid nature for bundles of the stem. This is confirmed here (Figs. 11, 12bc). As seen with SEM, the tracheids of Dracaena stems prove to have porose membranes in the scalariform pits on tracheid end walls (Fig. 11d). Cheadle (1942) reported scalariform perforation plates in leaves of Dracaena, and cited this as an exception to the root-stem--leaf sequence of vessel progression within a plant. However, the supposed vessel elements of Dracaena actually have porose pit membranes (Fig. 11e) and should probably be called tracheids.

Roots ofmonocots mostly do not develop secondary bundles. Widened stem bases do occur in Aloe, Beaucarnea, Cordyline, and Yucca, but roots are continually initiated on these stem bases as they widen. As plants of these genera increase in size, the diameter of the roots may widen, however, but they still, as far as is known, consist only of primary tissues. This is also true in roots of such non-asparagalean groups as palms (Iriartea) and Pandanaceae, both of which form conspicuous prop roots. Roots of large diameter have more numerous alternating xylem and phloem poles surrounding a central pith.

Dracaena is an exception in that its roots produce secondary bundles, as noted by Tomlinson and Zimmermann (1969). Dracaena draco, the dragon tree, bears thick roots at the bases of stems, roots which increase in thickness over time. These roots contain secondary bundles (Fig. 12a). The limits of the primary root and the beginning of the zone of secondary bundles are indicated in Fig. 12a (top). At left in Fig. 12a, are embedded in extraxylary fibers (dense area of light gray). Special note should be taken that the xylem in the primary roots of Dracaena draco (and other species of Dracaena) consists of vessels (Fig. 12b-c) rather than tracheids. The secondary bundles (Fig. 12a, right) consist, on the contrary, wholly of tracheids, and are collateral, with phloem (p) facing toward the outer surface of the root.

Special note should also be taken of the comparative diameter of the vessels in the primary root and the tracheids in the secondary bundles. The tracheids in the secondary bundles are notably wide in diameter but also may be thick-walled (Fig. 12d-c). In other words, there is a compensation, by means of wide tracheids, for the fact that vessels are absent in secondary bundles. This correlation has not been noticed earlier, but is essential to understanding the physiology of conduction in dracaenoid roots. The stems in the monocots listed above are vesselless, so that there is no way in which vessels of roots could be connected to tracheids in stems: the adventitious nature of monocots prevents that. Dracaena is highly distinctive among monocots in that secondary growth in roots of Dracaena can be intercontinuous with monocot cambia in steins, and therefore formation of vessels than extend from stems into roots is a theoretical possibility, but one that has not been realized in the dracaenoids or any other monocots. The intercontinuity of wide tracheids formed in secondary stem and root bundles may be considered a reasonable substitute. The advantages of adventitious roots in monocots (Carlquist, 2009; see also the "valve" hypothesis below) are sufficiently great that adoption of vessels that extend from roots into stems as in woody angiosperms would be of marginal value.

The addition of secondary bundles to stems by means of a monocot cambium is a way of achieving greater stature; palms, which do not have a monocot cambium, have an alternative series of adaptations, considered later. Most of the non-palm arborescent monocots have addition of secondary bundles as a way of achieving taller stature. Ravenala and similar strelitzioid genera may be considered arborescent by some, or may be excluded from the arborescent category; they do not have monocot cambia.

Protoxylem Wall Microstructure

In most primary walls of monocot metaxylem, networks of primary wall cellulosic fibrils can be seen in preparations in which amorphous wall portions are sectioned away (e.g., Figs. 1g, 2c-e) or in which amorphous material is characteristically hydrolyzed (e.g., Figs. lf, 2a-b, 11d, e). In most monocots, however, there is probably no cellulosic network in the primary walls of protoxylem (Carlquist & Schneider, 2011), as illustrated by grasses. This may be correlated with rapid elongation and expansion of protoxylem tracheary elements. However, in protoxylem of some monocots, such as Zingiberales (Carlquist & Schneider, 2010a), cellulosic strands are revealed by SEM (Fig. 13). These fibrillar strands are best illustrated in sections that have cut tracheary elements open, leaving the fibrils intact, rather than in tracheary element surfaces that have been split apart by sectioning.

If limited amounts of wall material are cut away, as in Fig. 13a, a few strands may persist. It the primary wall is not sectioned, thick strands running perpendicularly to the helical bars of secondary wall material are visible (Fig. 13b-f). These strands are presumably primary wall material, but this has not been demonstrated conclusively. The major strands that extend across primary walls of protoxylem tracheary elements in Zingiberales often fade into a reticulate pattern. This is noticeable in Fig. 13b, d, e, and f, but not evident in Fig. 13c.

Microstructure of protoxylem is a topic that has as yet been little explored in any group of angiosperms. The significance of cellulosic fibrils in primary walls of zingiberalean protoxylem elements may relate to nature of expansion. In genera and families of this order, expansion of protoxylem may be slow and limited compared with rapid and extensive elongation of protoxylem in, say, grasses. The presence of a cellulosic network could delay indefinitely the collapse of protoxylem primary walls. Ideas such as these can be tested by comparative investigations.

Monocot Xylem in the Context of Phylogeny

Early workers in xylem evolution took pride in the fact that their ideas were developed independently of a phylogenetic tree of angiosperms (Bailey & Tupper, 1918; Turrill, 1942; Cheadle, 1942; Bailey, 1944; Tippo, 1946; Cheadle & Tucker, 1961). This attitude was defensible only because DNA-based trees, such as the one in Fig. 15, were not available to them. Indeed, if such trees had been in existence, working on xylem evolution with reference to molecular phylogeny would have been considered mandatory.

In fact, Bailey was disingenuous in downplaying the role of the natural system in the development of wood phylogeny. In a long series of papers (with such workers as Nast and Swamy), he avidly studied the "woody Ranales" (= woody basal angiosperms in current phylogenies, such as APG III, 2009). By studying such suspiciously "primitive families", he was aware of phylogenetic thinking, but curiously wished to distance himself from it, perhaps because the efforts to construct a natural system at that time (e.g., Bessey, 1915) involved so much guesswork and the arbitrary use of "dicta." Also, the natural systems proposed in much of the 20th century were diverse in many key details, and the lack of consensus made them less than useful to those interested in evolution of structural features.

What criteria did Bailey and his students use for phylogenetic purposes under these circumstances? Bailey and Tupper (1918) identified an evolutionary trend, visible in vascular plants as a whole, for shortening of fusiform cambial initials (monocots were not included in the survey, however). In vesselless woody groups, tracheid length could be employed as a way of approximating the length of fusiform cambial initials. In vessel-beating woody groups, vessel element length is an accurate indicator for fusiform cambial initial length (vessel elements do not increase in length appreciably compared to the length of the fusiform cambial initial from which they were derived).

Bailey and Tupper (1918) must have realized that tracheary element length by itself is not an indicator of phylogenetic progression away from a hypothetical ancestor. Table VI in Bailey and Tupper (1918) divides woody angiosperms into four groups based on character state changes in wood anatomical features: lateral wall pitting (beginning with scalariform, ending with alternate); and degree of border presence on pits of imperforate tracheary elements (fully bordered, ending with absence of borders). Bailey seems to be saying that morphological features can be used interchangeably with tracheary element length as phyletic indicators (see Tippo, 1946). Indeed, Bailey handed off these features to graduate students. Frost (1930a, b, 1931) detailed angularity of vessels as seen in transection; end wall angle of vessel elements; number of bars per perforation plate; and lateral wall pitting of vessels. Kribs analyzed degrees and kinds of aggregation of axial parenchyma (1935) and change in ray histology (1937).

To Vernon Cheadle fell the task of determining how xylem evolved in monocots. Cheadle (1942) considered that longer vessel elements should be considered a symplesiomorphic ("primitive") character state, and retained that view (Cheadle & Tucker, 1961) and never questioned it. In fact, this assumption is fallacious, because monocots do not have vascular cambia, and therefore do not have fusiform cambial initials. In monocots, vessel element length is governed by other factors, such as degree of organ elongation, activity of basal meristems, size of plant, etc. The data of Klotz (1977) show that climbing palms with long internodes have longer vessel elements than do upright palms, for example.

Now that we have molecular-based trees inclusive of many families for monocots (Fig. 14), we can see that other assumptions made by Cheadle (1942; Cheadle & Tucker, 1961) are unfounded. Cheadle thought that vessels originated independently in monocotyledons and dicotyledons, but that was at a time when monocots and dicots were thought to represent the products of the first forking of the angiosperm tree. That idea was abandoned with the first global tree of angiosperms (Chase et al., 1993) and all subsequent trees. "Basal angiosperms" are now regarded as the ancestral group in which monocots are nested. All of the basal angiosperms have vessels except for Amborellaceae, Ceratophyllaceae, Nymphaeales, and Winteraceae. These groups are probably not the direct ancestors of monocots, which seem rooted more closely to the vessel-beating groups Chloranthales and Piperales (Carlquist, 1992a, b, 2009) where structural resemblances are concerned. Winteraceae are basal angiosperms, but not close to the origin of monocots, and very likely are secondarily vesselless (Young, 1981; Chase et al., 1993; Soltis et al., 2000).

One should mention that woody non-monocot angiosperms all have vascular cambia, and that cambial loss is one of the earliest character state changes, if not the earliest, that led to monocots. The loss of cambium is well illustrated in Houttuynia of the Saururaceae (Carlquist, 2009), although that genus is not ancestral to monocots.

The studies of Bierhorst and Zamora (1965) show that in families and species from 165 angiosperms (including basal angiosperms, sensu APG III, 2009), primary xylem contains vessels in all of the species they studied. Bierhorst and Zamora (1965) report tracheids as well as vessels in protoxylem of many of the species they studied, and note a trend of specialization, expressing itself in the earlier ontogenetic appearance of advanced features and the elimination of primitive ones. The only families in which Bierhorst and Zamora note some tracheids (along with vessel elements with scalariform perforation plates) in metaxylem are Aquifoliaceae (Ilex), Buxaceae (Pachysandra), Caprifoliaceae (Weigela), Cornaceae (Comus), Cunoniaceae (Spiraeanthemum), and Ericaceae (Gaultheria). The omission of Chloranthaceae from these studies is regrettable, because one might have found that primary xylem of Sarcandra stems lacks vessels, as suggested by the results of Bailey and Swamy (1950). Sarcandra develops discernable vessels only in secondary xylem of roots or caudices (Carlquist, 1987). The primary xylem in Winteraceae and Trochodendraceae is evidently vesselless, also (Carlquist, 2009). Some of the species studied by Bierhorst and Zamora (1965) might have proved to have pit membranes in end walls of vessel elements, if they had been able to undertake SEM studies of sections instead of light microscope studies of macerations. Primary xylem is mentioned here because it was alleged by Bailey (1944) to be a sort of refuge for primitive features, so that if monocots and non-monocot angiosperms independently acquired vessels, we might expect to see all-tracheid primary xylem with vessel-bearing secondary xylem. That is evidently not always the case.

The available data and molecular trees now produced suggest that Cheadle's contention (Cheadle, 1942; Cheadle & Tucker, 1961) that monocotyledons are primitively vesselless and that vessels originated independently in monocotyledons and non-monocot angiosperms should be questioned. We can no longer accept the dictum of Bailey (1944): "The independent origins and specializations of vessels in monocotyledons and dicotyledons clearly indicate that if the angiosperms are monophyletic, the monocotyledons must have diverged from the dicotyledons before the acquisition of vessels by their common ancestors. This renders untenable all suggestions for deriving monocotyledons from vessel-bearing dicotyledons or vice versa." Bailey's statement is incorrect now that we have information from DNA-based phylogenetic trees. He also fails to take into account the profound differences related to growth form. Sympodial angiosperms with adventitious roots inevitably have patterns of vessel evolution different from those seen in monopodial woody angiosperms with taproots.

Symplesiomorphy in Monocot Xylem: Can we Find It?

If we compare the molecular-based tree of Fig. 15 to what is known about the xylem of these families, do we find concordance or discordance? If we find discordance, why?

Heteromycotrophic monocots are apparently vesselless, although we do not have complete information on all of them (notably heteromycotrophic Burmanniaceae and Orchidaceae). The heteromycotrophic families (signified by circles at tips of branches in Fig. 15) do not group closely. Rather, they are homoplasic, as the tree in Fig. 15 suggests (this would be even more evident if heteromycotrophic orchids were plotted). Although one family (Petrosaviaceae) is an early-diverging branch of monocots, one genus (Japanolirion) is autotrophic. We can safely conclude that vessellessness in heteromycotrophic monocots (still insufficiently studied) is secondary. This is instructive, in that these monocots can serve as an example of how secondary vessellessness can occur.

After Acorales (= the genus Acorus),which is the sister to the remaining monocots and is discussed separately below, the next node leads to Alismatales. Araceae do have vessels in roots (Carlquist & Schneider, 1998; Schneider & Carlquist, 1998), although some pit membrane remnants can be found in some perforation plates. No convincing evidence for presence of vessels in stems of Araceae has been presented.

The clade that contains Alismatales (Fig. 15) can be characterized as consisting mostly of aquatics. Notable is the fact that the submersed aquatics in the order (Aponogetonaceae, Hydrocharitaceae, Najadaceae, Ruppiaceae, Zannichelliaceae, Zosteraceae) lack vessels throughout the plant. Submersed aquatics may have adapted to that habit/habitat recently, but the likelihood is that the earliest monocots were not submersed aquatics and that vessels may have been present in the roots. One notes that in Fig. 15, the family Tofieldiaceae is the sister to the remaining Alismatales. To be sure, this sampling is less than optimal. However, Tofieldiaceae and Alismataceae, although characteristic of marshy habitats (ranging from savannah seeps to ponds), have vessels in roots. To be sure, one must always keep in mind that the xylem of plants is likely to relate to the present-day ecology of the plant, and not represent relictual conditions. Imagining a symplesiomorphic status for vessellessness in the submersed families of Alismatales would require nonparsimonious character state reversions. Much more likely is the idea that presence of vessels in roots of aquatic monocots is symplesiomorphic, and is related to occupancy of habitats in which roots experience some degree of fluctuation of moisture availability, making vessels advantageous. Submersed aquatics have developed intricate (and diverse) means of coping with low oxygen levels in water, mechanisms that would have to be developed and then lost again if submersed aquatics were to represent the ancient monocot habitat. Absence of vessels in roots, of the submersed aquatic monocots is, therefore, probably apomorphic, representing secondary vessellessness.

Autotrophic habits and nonaquatic habitats are characteristic of the monocots near early nodes of the tree in Fig. 15 and other trees that have been proposed (Davis et al., 2004; APG III, 2009). Such habitats are characteristic of monocots, in which one sees, with few exceptions, presence of vessels in roots, but lack of vessels in stems and leaves (Fahn, 1954; Cheadle, 1963, 1968; Cheadle & Kosakai, 1971; Wagner, 1977; see also original data above). According to Cheadle (1968), Campynemataceae are possibly totally devoid of vessels. Original data on Lapageria (Philesiaceae) given above (Fig. 8a-b) suggests that it may fall in that category also. A few other instances may be found in Liliales when they have been more intensively investigated. The SEM data on Asteliaceae (Carlquist & Schneider, 201 Ob) and Orchidaceae (Carlquist & Schneider, 2006; see also Fig. 3) are persuasive that some Asparagales have no vessels. In Dioscoreales, vessel presence has not yet been clearly established throughout Taccaceae (Fahn, 1954; Cheadle, 1968; Wagner, 1977). All of the Liliales and Dioscoreales listed occur in highly mesic habitats, but that does not necessarily indicate that the genera and families just cited are relictual in lacking vessels in roots (as well as in stems and leaves). Secondary vessellessness has been claimed for Winteraceae and Trochodendraceae (Young, 1981, and subsequent authors), and those two families are limited, as are the monocots just mentioned, to highly mesic localities in which there is little fluctuation in water availability.

However, to the above, one can add families and genera that lack vessels in stems and roots and have very "primitive" vessels (long scalariform perforation plates) in roots. Cyclanthaceae, Pandanaceae, and Typhaceae have been highlighted above in this regard because earlier reports suggested that these three families have vessels throughout the plant. Families in which vessels with long scalariform perforation plates occur in roots whereas stems and leaves have only tracheids include Araceae (Keating, 2003), Costaceae, Hanguanaceae, Heliconiaceae, Hypoxidaceae, Melanthiaceae, Petermanniaceae, Ruscaceae, Trilliaceae, Zingiberaceae, and a number of genera of hyacinthoid Liliaceae (Wagner, 1977).. Numerous genera from Orchidaceae and various other families could be added to this list. In other words, this appears to be a widespread condition in the earlier-departing clades of monocots (as schematized in Fig. 15: orders from Acorales upward to Asparagales). If one views the distribution of vesselless or near vesselless genera, they do not appear to be the earliest branches in their respective clades in Fig. 15. If one were to hypothesize vessellessness as symplesiomorphic for monocots, one would have to account for multiple instances of vessel acquisition, if the tree of Fig. 15 is tenable.

One can hypothesize that the presence of long scalariform perforation plates in roots combined with only tracheids in stems and leaves is not only symplesiomorphic for monocots as a whole, but also that it has adaptive significance. Roots tend to have wider vessels than stems in woody dicots (Patel, 1965), and if this is true for monocots as well, then vessels are more likely to occur in roots than in stems of monocots. Adventitious roots by their very nature experience more fluctuation in water availability than do taproots, so the presence of vessels in roots of monocots is understandable.

If one views the ecology of the monocots with this xylem formula (vessels with scalariform perforation plates in roots, only tracheids in stems), one sees that they mostly inhabit highly mesic localities. Some genera on this list have mitigating conditions, such as succulence, that permit them to function as "temporary mesophytes" (e.g., many orchids; Hyacinthaceae). The hypothesis most in line with molecular trees of monocots, knowledge of tracheary element morphology, and ecology is twofold. Genera with this formula have a symplesiomorphic xylem condition, and they have had unbroken occupation of mesic habitats. Departures from this formula must have taken place homoplasically, and such divergences represent numerous clades that have adapted to progressively more seasonal conditions. These departures represent tradeoffs between conductive efficiency (vessels with simple perforation plates) and safety (an all-tracheid condition). The patterns of xylary apomorphies in monocot xylem are numerous and should be traced on a family-by-family basis. Commelinales and Poales are not covered to any appreciable extent in the present paper, because they are crown groups that have already attained extensive vessel presence throughout the plant--a feature deserving of ecophysiological study, notably different from the presence of all-tracheid systems in monocots.

The Role of Ontogeny and Cell Size in Vessel Presence

The simplest explanation for presence or absence of pit membranes in a vessel end wall is a developmental one. The pit membranes are swept away by the conductive stream because they have an insufficient cellulosic network to resist the effects of the flow. The nature of the cellulosic network is, presumably a feature embedded in the genetics and development of the vessel elements. As yet, we do not have comparative tracking of cellulosic network presence in pit membranes or stages in its loss as vessel elements mature. Secondary vessellessness may be achieved by relatively minor changes in the cellulosic components of the pit membrane. If cellulosic fibrils are present in pit membranes of tracheid end walls, pit membranes may be retained as a result of gene action, resulting in absence of lysis of the pit membrane, rather than (as is typical for perforations in vessel elements), swept away in the flow of xylem sap. Such possibilities are developmentally simple and plausible causes of retention of the tracheidlike characteristics of a xylem cell, and should be considered before other possibilities are entertained.

There are, however, other ways in which secondary vessellessness may occur. Klotz (1977) showed that in palms, imperforate tracheary elements are present in all species in early metaxylem, whereas vessels occur in late metaxylem. Could this lead, phylogenetically, to an all-tracheid system if production of late metaxylem was suppressed? Theoretically, yes, but definitive demonstrations of such shifts may be difficult. Nevertheless, some examples are suggestive, and are worthy of discussion.

The clearest examples of this trend are in the submersed aquatics of the Alismatales such as Aponogetonaceae or Zosteraceae, in which vessels may have been lost simply because so little xylem is produced. Something like this may have happened in commelinalean family of submersed aquatics, Mayacaceae, also. Mayacaceae have long scalariform perforation plates in roots, tracheids only in stems and leaves. Mayacaceae are nested within Commelinales that have more "specialized" xylem (see Fig. 15), a contradiction of a dictum by Cheadle (see next section). Mayacaceae may merely be forming protoxylem and early metaxylem, in which tracheids and scalariform perforation plates are to be expected.

Also possible examples of this may be found in Philesiaceae (Lapageria), Taccaceae, and Campynemataceae. They may be forming no "late" metaxylem as defined by the presence of wider tracheary elements. This could also be true in the palm Phytelephas and its close relative Ammandra, which lack vessels in stems, and have relatively narrow metaxylem elements (Klotz, 1977). Perhaps we should think in terms of narrowing of tracheary elements rather than disappearance of vessels. Narrrower tracheary elements in protoxylem and earlier formed metaxylem are more likely to ,be tracheids than vessels, and more likely to have scalariform perforation plates than simple ones as compared to metaxylem. This idea was enunciated by Bailey (1944) who thought of the primary xylem as a refuge for "primitive" xylem characteristics in woody angiosperms. Bierhorst and Zamora (1965) found evidence to support this idea in their study of primary xylem, as did Cheadle (1968) in Haemodoraceae.

Monocot bundles may be considered juvenile in comparison with those of angiosperms capable of vascular cambial activity. The developmental sequence can therefore be regarded as foreshortened, or juvenilistic. Monocot bundles have been so regarded in a study that places xylems of angiosperms within a developmental framework (Carlquist, 2009). That study was conceived in terms of activity of the vascular cambium. However, one may, by extension, add ontogenetic changes within a bundle that has no vascular cambium. Monocot bundles that do not proceed all the way to typical late metaxylem patterns can thus be called juvenilistic. Typhaceae are mentioned above as an example of how early metaxylem tracheary elements of roots have scalariform end walls that retain pit membranes, whereas late metaxylem tracheary elements are genuine vessel elements that lack pit membranes (except as fragmentary remnants) in perforation plates. Evolutionary deletion of late metaxylem in such a clade could result in secondary vessellessness.

In another perspective, one may consider that vascular bundles of monocots can exhibit various degrees of dimorphism. In palms, for example, the late metaxylem vessels are much larger (and more likely to have fewer bars on perforation plates) than the early metaxylem, and early metaxylem apparently always contains tracheids whereas late metaxylem lacks tracheids (Klotz, 1977). Dimorphism between late metaxylem vessels and protoxylem + early metaxylem vessels is also familiar in the transectional configurations one sees in grass vascular bundles (Metcalfe, 1960). In this perspective, abrupt differences between early and late metaxylem are undoubtedly mediated by hormonal action.

Although both of the above perspectives seem valid, one is still left with the question as to why these ontogenetic progressions occur, and are foreshortened or abruptly changed. Morphological goals are reached not as fulfillments of inexorable changes, but in response to functional value in the environment. There seems little doubt that wide vessels, as in palms, are formed in response to the increase in conductive capability by the fourth power of the increase in vessel diameter (the Hagen-Poiseuille equation, Tyree & Zimmermann, 2002). Such wide vessels are, however, potentially vulnerable, because wider vessels embolize more readily than narrower ones, as indicated by vessel diameter changes in growth rings (Carlquist, 1980), and as can be proved experimentally (Hargrave et al., 1994). The sheathing of wide vessels in palms by parenchyma (Tyree & Zimmermann, 2002) suggests that parenchyma may form a system that counteracts vulnerability to some extent. Root pressure (which is controlled by parenchyma in ways not fully demonstrated yet) may also play a rote in countering vulnerability in wide vessels such as those of palms (Davis, 1961). Angiosperm tracheids, on the other hand, are quite resistant to spread of embolisms from one cell to the next by having pit membranes that resist air bubble transfer by having small pore size. Increasing thickness of pit membranes reinforces this capability (Jansen et al., 2009). Tracheids also generally have small diameter, and thus, as the Hagen-Poiseuille equation tells us, are less efficient in conduction (conifer tracheids do not conform to this, for reasons stated by Pitterman et al., 2005), but we are dealing here with angiosperms, in which coniferous kinds of margo-torus pit membrane structure has never evolved in tracheids. There are instances oftori or pseudotori, which may seal off bordered pits when pressure differences among cells develop, in woody dicots such as Oleaceae, Ericaceae, Thymeleaceae, and Ulmaceae (e.g., Dute & Rushing, 1987; Rabaey et al., 2006). These do not have the conductive advantage possessed by the margo in conifer tracheid pit membranes. In any case, monocots are not known to have tori or pseudotori.

The shift from protoxylem tracheary elements to late metaxylem elements has usually been seen in structural terms, from extensible wall patterns (annular, helical) to non-extensible pitted patterns. This common textbook story does not take into account a shift from low conductive abilities combined with conductive safety (protoxylem, early metaxylem) to high conductive abilities combined with increased vulnerability (late metaxylem). The ways in which such xylem patterns relate to the physiology and ecology of a species are left unexplored in favor of the more easily described wall patterns, readily shown with light microscopy. Monocots show an organographic balance between conductive efficiency and conductive safety. This balance is not possible in woody angiosperms because the vascular cambium produces continuity from root to shoot. This resulting vascular continuity lacks the valve (or "rectifier") feature that adventitious roots supply (see below). The conductive safety/conductive efficiency balance can be regulated in monocots by production of roots of finite (often very short) duration on stems of longer duration. It can also be accomplished by curtailment of or sudden shifts in the protoxylem/metaxylem progression. Thus, terrestrial monocots with very narrow tracheary elements, such as Lapageria or Campynema, can manage without vessels or with very tracheidlike vessels because they have mesic ecology matched with low transpiration rates, and can satisfy their water economy requirements with xylem low in conductive efficiency. Xylem formulations should always be viewed within ecological and physiological contexts. To view them merely as externalizations of degrees of evolutionary progress eliminates consideration of the forces that drive change in anatomical patterns and robs them of their significance.

Ecological Iterations: A Key to Paradoxical Distributions of Xylem Character States

Cheadle (1942) thought of xylem formulas in monocots as representing levels or grades of specialization, and he developed numerical ratings to record degree of advancement for any taxonomic group. His five-point scale is given above under Historical Perspectives. Rating evolutionary advancement is a data sink: it is condensed from real and valid data, but because it produces generalizations, it cannot be used to yield new perceptions or conclusions about particular species: it cannot tell anything about how these species and clades evolved, in relationship to what factors.

Symplesiomorphic xylem characters should be expected for monocots that have had unbroken histories of occupancy of mesic habitats, and which therefore exhibit no "ecological iteration" (shift in habitat preference). Long scalariform perforation plates in vessels of roots, combined with only tracheids elsewhere in the plant body, characterize such diverse groups as Campynemataceae (moist rock outcrops in New Caledonia and Tasmania); Lapageria (moist forest in southern Chile), Petermannia (moist forest in Queensland and New South Wales), and Cyclanthaceae (understory of wet neotropical forests). These would correlate with their position as early branchings within clades of monocots (Fig. 14). There is no reason to believe that scalariform perforation plates have been secondarily derived from simple plates by some kind of morphological reversion. Clades with simple perforation plates in xylem can radiate into less seasonal habitats. For example, grasses can occupy extremely wet areas, despite the fact that their xylem (vessels with simple perforation plates, throughout the plant body) probably evolved in response to highly seasonal conditions, drawing water from shallow soil depths. Thus, scalariform perforation plates in roots often do represent a symplesiomorphic feature--but with numerous cautions mentioned above. The genera and families just cited have distribution patterns that correspond to ancient land areas.

However, good dispersal may permit a monocot with such an antique xylem formulation to reach geologically new and highly disjunct land areas--as long as they are ecologically suitable. This is true of Astelia, for example, which has apparently dispersed from areas like New Zealand and Australia to Islands as distant as Reunion in the Indian Ocean and Tahiti and the Hawaiian Islands in the Pacific. Astelia has baccate fruits with small seeds suited to bird dispersal. Corresponding to its symplesiomorphic xylem features, it occupies consistently moist forests or similarly mesic areas. Typha, which has preferences for sunny ditches and muddy depressions and has a xylem configuration similar to that of Astelia, has a very wide boreal distribution because of its tiny windborne seeds. Thus, one should not expect symplesiomorphic xylem to correlate with extent of ancient geological areas, although in some instances it does.

Rapid evolution into highly seasonal habitats is certainly characteristic of many monocot clades: there are always more evolutionary opportunities in environments with more fluctuation in temperature and precipitation, because extinction is likely to be greater in more extreme habitats and therefore, niches are more readily available. This triggers the question: can monocot clades that in their contemporary species have more "specialized" (= vessels in stems and/or leaves in those in roots; perforation plates with few bars or simple) xylem branch off early, while "crown groups" retain more "primitive" configurations? The answer is yes, but Cheadle negates this possibility. He argued that xylem specialization was identical to phylogenetic specialization, and that therefore a group with "specialized" xylem could not be ancestral to one with "primitive xylem." One example is the apostasioid orchids (Apostasiaceae of some authors), which molecular trees uniformly show branching off at the base of the clade leading to the other orchid subfamilies (Davis et al., 2004; Fig. 14). Cheadle and Tucker (1961) say, "Apostasiaceae ... cannot have been the origin of Orchidales". Cheadle consistently negated the possibility that plants with simple perforation plates might be the survivors of a line with numerous symplesiomorphic characters. Today, we would say the "breakouts" favoring rapid evolution of simple perforation plates can occur in any clade. In fact, xylem designed for water economy in more extreme habitats would favor long-term survivorship in a line in which, say, floral characters are symplesiomorphic, as they are in the apostasioids (Judd et al., 1993; Kocyan et al., 2004) Aspostasioids have carried along symplesiomorphic DNA sequences in plants that have evolved to cope with marked fluctuation in moisture availability.

An even more striking example can be found in Borya (Boryaceae), mentioned above (Fig. 9) is Borya, in which vessels with few bars on perforation plates occur in stems. This xylem formulation correlates with the "resurrection plant" habit of Borya, which lives on granite outcrops that dry quickly after winter rains. Borya is near-basal in the Asparagales clade, nearly all of the genera of which lack vessels in stems. Interestingly, one of the few exceptions to that description of Asparagales is Sisyrhynchium (Iridaceae), which has vessels in stems and leaves (Cheadle, 1963) and maintains foliage in summer-dry habitats of the southwestern U.S. Boryaceae probably belong to a group that once included genera with symplesiomorphic xylem features, genera that are now extinct. This is not an improbable scenario, and other examples can be cited within monocots. In the tree of Fig. 15, Arecaceae (vessels throughout the plant, except in a few non-basal genera) is a sister family to Zingiberales (vessels in roots only except for Cannaceae, Marantaceae) plus Commelinales (vessels various). Velloziaceae (vessels with simple perforation plates in roots) plus Triuridaceae (a vesselless heteromycotroph) bear a sister relationship to the remaining Pandanales (Fig. 14). All of the remaining Pandanales have more symplesiomorphic xylem configurations (vessels with long scalariform perforation plates in roots, but probably no vessels in stems or leaves). Velloziaceae have adapted to tropical savannah-like habitats with seasonal fluctuation in water availability that corresponds to presence of simple perforation plates in root vessels. Rhizogonaceae, with simple perforation plates in roots (Fahn, 1954), is probably a sister group to Philesiaceae (either tracheids only, or possibly very long scalariform perforation plates in roots). The pairs of close families just cited show how one family of a pair may have adapted to highly seasonal habitats while its sister group continued an unbroken occupancy of mesic habitats.

Probable loss of metaxylem vessels phylogenetically in submersed aquatics make some "crown groups" seem to have more "primitive" xylem than they do. For example, Mayacaceae, with vessels with long scalariform perforation plates in roots and stems, is a "crown group" nested among non-aquatic commelinalean groups that have vessels with fewer bars per perforation plates, according to the data of Tomlinson (1969). Likewise, the submersed aquatics of Alismatales lack vessels, although Araceae, with vessels in roots, are sister to the Alismatales. All of these examples underline the principle that xylem designs are adaptive in contemporary situations, and we should not look to them as reliable sources of phylogenetic history.

Thus, there are at least two possible scenarios for why "specialized" xylem may appear in groups basal in particular clades. One can find these scenarios in nonmonocot angiosperms also. Within Ranuneulales, Papaveraceae (vessels with simple perforation plates throughout the plant, even in primary xylem: Bierhorst & Zamora, 1965; Carlquist & Zona, 1988) is sister to a group of families that includes Lardizabalaceae and Ranunculaceae. Decaisnea of the Lardizabalaceae has long scalariform perforation plates (Carlquist, 1984b). Papaveraceae have probably radiated into highly seasonal habitats, and surviving Papaveraceae show no traces of earlier stages of this radiation.

In general, the more symplesiomorphic xylem configurations are more abundant in more basal positions within clades. This would verify Cheadle's generalization, with exceptions like the above. However, there is one major caveat. The most symplesiomorphic xylem condition in monocots, according to Cheadle, is an all-tracheid condition. According to modern phylogenies (e.g., APG III, 2009), monocots represent a clade branching from basal angiosperms close to the departure point of Chloranthales. Earlier branchings in the basal angiosperms include the outgroups shown at the bottom of Fig. 14, according to all modern molecular trees.

A quotation from Bailey (1944) is appropriate to show how far we have advanced in our thinking, and how much paradigms of monocot xylem must be changed:

"The independent origins and specialization of vessels in monocotyledons and dicotyledons clearly indicate that if the angiosperms are monophyletic, the monocotyledons must have diverged from the dicotyledons before the acquisition of vessels by their common ancestors. This renders untenable all suggestions for deriving monocotyledons from vessel-beating dicotyledons or vice versa. Furthermore, the highly specialized structure of the xylem throughout both stems and roots of herbaceous dicotyledons, not only affords conclusive supplementary evidence of the derivation of herbaceous from arboreal or fruticose dicotyledons, but also is an insuperable barrier to the derivation of monocotyledons from herbaceous dicotyledons."

Newer information has made Bailey's statement untenable. Evidence from DNA-based phylogenies refutes Bailey's thinking, as all of the global angiosperm trees produced since Chase et al. (1993) show. In that respect, Bailey was a victim of his time. However, the understandings of Bailey and of Cheadle were seriously limited by methodological procedures. They never related xylem structure to ecology of species, which is especially curious considering that Cheadle field-collected much of his material. Both viewed xylem as an inexorable progression (the stages of which could be given numerical ratings). They did not correlate xylem with habit. They did not consider relevant work in conductive physiology. They did not study developmental sequences and changes in tracheary element diameter within a xylem sample. And they did not consider the role of hormonal changes and translocation which were beginning to be appreciated in their time. Ultimately, of course, the governance of hormonal change by gene action must be included in the evolutionary picture. However, the central point is that in order to understand evolution of xylem structure, we cannot exclude information from habit, ecology and physiology.

Evidence for Terrestrial Versus Aquatic Origin of Monocots

Are monocots as a whole ancestrally aquatic or terrestrial? Current interpretations indicate that angiosperms as a whole are ancestrally sympodial, a growth form that is very frequently associated with adventitious roots, and that taproots and monopodial structures are probably apomorphies within angiosperms (Carlquist, 2009). To be sure, monopodial growth forms appear to have originated early in angiosperms (Carlquist, 2009).

Certainly monocots are better adapted to habitats that are wet for prolonged periods, because taproots, characteristics of woody angiosperms, are not effective in inundated habitats low in oxygen. The prostrate sympodial habit common in most monocot clades has the advantage of being able to spread laterally over territory, whereas a monopodial eudicot is limited to a narrowly limited piece of ground. Lateral spreading, accompanied by development of adventitious roots, is certainly adaptive in moist habitats, but more lateral spread implies capability of dealing with various degrees of moisture availability close to the ground surface, and therefore presence of vessels in roots can be hypothesized as advantageous.

If one were to hypothesize a submersed aquatic, or an aquatic with submersed stems but leaves emergent above the water surface as an ancestral habit/.habitat in monocots, non-parsimonious probabilities emerge. One has to imagine that somehow such submersed aquatics acquired methods for ventilating roots and stems in low oxygen habitats, then lost these mechanisms as most of the descendents moved onto terrestrial or occasionally inundated habitats. Because of the limitations imposed by low oxygen (and often low nitrogen or other nutrients) in standing water, the number of species adapted to such habitats is necessarily small. The number and area of habitats with a range of moisture availability ranging from seasonally inundated to sometimes dry is relatively large, on the contrary. To imagine a submersed aquatic origin for monocots, one would have to imagine them entering a very difficult, limited habitat first, then spreading to habitats for which varied xylem and parenchyma histology is suitable. To be sure, there are some apparently ancient groups that lack clearly defined vessels, such as Nymphaeales and Acorales, but these may have survived precisely because they long ago entered minimally contested habitats.

In addition, if one hypothesizes aquatic origin for monocots, one must imagine multiple origins of vessels, and multiple events in which vessels specialized in terms of organography and morphology. These multiple events would have to be imagined as having parallel outcomes instead of diverse ones. One would have to imagine that vessels originated from tracheids, always yielding the same scalariform end wall pattern, always simplifying the perforation plate in the same way.

Vesselless Stems and Leaves: Why so Common in Monocots?

Cheadle (1942, 1943a, b) posited that vessels originated first in roots and then spread, in the course of evolution to stems, inflorescence axes, and leaves successively. Accepting that this is the as a generalization, one could look at it in the reverse perspective: progressive loss of all-tracheid conditions. Monocot species with stems and leaves that lack vessels are perhaps as numerous as monocot species with vessel-bearing stems and leaves, although some conspicuous and speciose families, such as Cyperaceae and Poaceae do fall in the latter group (see Carlquist, 1975, p. 106). In fact, acquisition of vessels in Cyperaceae and Poaceae may well have helped accelerate their evolution. They deal with the environment in ways quite different from those of, say, the asparagalean families. Vessels throughout a plant body characterize most non-monocot angiosperms, so the physiology of all-tracheid systems has been neglected.

One explanation for maintenance of an all-tracheid xylem in stems and leaves has to do with the sympodial habit of most monocots, in which roots are adventitious. Developmentally, vessels in adventitious roots cannot connect with vessels in a stem. This disjunction between stem and root xylem necessarily produces differential opportunities for roots and stems. The numerous species of bulbous monocots have exploited this disjunction. Roots are rather ephemeral, and have xylem with (mostly) simple perforation plates in vessels suited for rapid conduction during seasons when moisture is available and frost does not exist. This formula is combined with all-tracheid xylem in leaf bases and stems, which are perennial. The leaf bases and stems can thereby persist through the dry season, the water columns of the tracheids unlikely to embolize, the parenchyma of the leaves serving for water and photosynthate storage. In many bulbous monocots, the upper photosynthetic portions of the leaves are succulent, and can persist well into the dry season (Calochortus, for example), attenuating the growing and flowering season.

Similarly, orchid stems are different from orchid roots in the way they deal with water economy--a fact that is doubtless basic to the amazing speciation of epiphytic orchids with their succulent pseudobulbs and leaves. Orchids also have mechanisms such as C4 photosynthesis (Silvera et al., 2010) and thick cuticles that provide ways of dealing with the special water and light economy in the epiphytic mode of existence.

Succulence and a suite of other features (Nobel & Hartsock, 1978, Woodhouse et al., 1980; Nobel, 1988) permit Agave to combine large leaf size with an all-tracheid xylem configuration in desert environments. Similar considerations apply to Yucca (Smith et al., 1983).

Some monocots with vesselless stems and leaves can become trees. Aloe dichotoma, Beaucarnea recurvata, Cordyline australis, Dracaena draco, and Yucca brevifolia are among a number of species that could be mentioned in this regard. The degree of arborescence of these species is not unlimited. All of them except Dracaena have adventitious roots that apparently lack monocot cambia and are formed on a widened base, or in some other fashion (Yucca brevifolia can form underground stolons). Adventitious vessel-bearing roots can in these be continually formed to supply the vesselless xylem of stems. The monocot cambium characterizes stems of all of these arborescent forms, and thereby provides a way of increasing the stem vasculature. There is, at the same time, an increase in either diameter or number of the adventitious roots. All of the arborescent monocots may be said to have succulent stems, and some of them have succulent leaves (Aloe most obviously). The leaf boundary layers of arborescent monocots should be investigated more fully, because they, like the transpiration reduction methods of orchids, can very likely be correlated with the advantages and limitations of a vesselless system. In fact, the work of Smith et al. (1983) provides considerable illumination on the ecophysiology of Yucca brevifolia and how it survives with an all-tracheid system in a desert environment..

Dracaena draco and other species of Dracaena have monocot cambium in roots, as well as in stems. This means that except for the metaxylem that functions near the tips of roots, Dracaena has a vesselless conductive system. As shown above, the large diameter of tracheids formed from monocot cambia in roots of D. draco probably compensates for the lack of vessels in the secondary bundles of stems and roots. Monocot cambia always produce only tracheids in the secondary bundles. We need more studies on arborescent monocots that link anatomy with function, because arborescent monocots are, in fact, relatively little known. Lack of such studies is probably the result of lack of commercial importance of arborescent monocots, and the fact that most of them do not grow near major academic institutions.

Habit and habitat correlate very closely with xylem anatomy, and xylem anatomy should be studied in this regard. The above observations, which may be regarded merely as obvious, are illustrative of opportunities for study of all-tracheid systems in stems and leaves of monocots in relation to habit and habitat. The all-tracheid xylem found in stems and leaves of so many monocots is compatible with management of moderate fluctuations in conductive rates, but there are many physiological mechanisms for maintaining lower conductive flux while maintaining plant size and variety in leaf construction.

Is the vesselless condition in stems and leaves of the vast majority of Arales, Alismatales, and Asparagales primitive or secondary? Parsimony would dictate that multiple instances of loss of vessel-bearing metaxylem in these orders are unlikely to have happened, and that if trees such as those of Davis et al. (2004) and that of Fig. 15 are valid, vessellessness in monocot stems and leaves is a symplesiomorphy.

Palms: Unique in Habit and Xylem

Palms do not have a monocot cambium, a salient fact that separates them from other arboreal monocots. The widened bases of palm trunks are not the result of lateral meristem activity, they are masses of accumulated roots. The fact that palms do not have monocot cambia means that they have developed a series of xylem strategies different from those of the arboreal monocots with monocot cambia (Kingia of the Dasypogonaceae has very little if any monocot cambium activity, and offers some comparisons with palms).

Palms often have vessels with simple perforation plates in roots, scalariform or simple perforation plates in stem vessels with scalariform perforation plates in leaves (Tomlinson, 1961; Klotz, 1977; Wagner, 1977). There are a few exceptions to this pattern, and these exceptions prove unusually interesting.

Nypa is a prostrate palm of estuarine margins. Its stems are dichotomously branched and prostrate. The root vessels have long scalariform perforation plates in roots and stems (Klotz, 1977). Leaves have tracheids only. This xylem might be expected if Nypa (one species, N. fruticans) were phylogenetically a basal branch of Arecaceae, but it is not. In fact, the calamoid palms appear to be sister to the remaining Arecaceae (Baker et al., 2009). Calamoids (the genus Calamus) have simple perforation plates in roots and stems, and a mix of simple and scalariform perforation plates in leaves (Tomlinson, 1961; Klotz, 1977)--a xylem conformation one would call apomorphic within the family. This paradox may be explainable if we take into account habit and ecology. Lianas tend to have wide vessels, in almost any angiosperm group (Carlquist, 1975). Calamus is lianoid and has wide vessels (those of Calamus stems 100-460 [micro]m, with simple perforation plates: Klotz, 1977) in almost any angiosperm group, and wider vessels tend to have simple perforation plates (Carlquist, 1975). There are also nonlianoid calamoid palms, so that we need not postulate the lianoid habit as symplesiomorphic in palms. The Cheadle dicta (Cheadle & Tucker, 1961) would claim that nypoid and chamaedoroid palms could not have been derived from calamoid palms. Cheadle did not take into account such factors as habit and vessel diameter, merely perforation plate morphology, in attempting to establish status of a particular group on a sort of phylogenetic ladder.

Nypa has comparatively narrow late metaxylem vessels in roots (120-130 [micro]m in diameter) and stems (80-120 [micro]m in diameter). Klotz (1977) somewhat whimsically wonders whether or not we should consider Nypa as a xerophyte because it grows along estuaries which may vary in saltiness. The answer is no, because availability of water, even if saline, is more important than mineral concentration. In Caryophyllales, Frankeniaceae have xeromorphic xylem. Frankeniaceae have relatively shallow root systems that draw from soil depths that are not only salty but often dry. In Frankeniaceae narrow vessels are in groups, an indication of xeromorphy (Carlquist 1984a, 2010). The neighboring family Tamaricaceae taps deeper levels that are perpetually moist, if saline, and thus has relatively mesomorphic xylem (vessels wide, solitary). Tamaricaceae are therefore "hydrohalophytes" (Carlquist, 2010).

The phytelephoid ('vegetable ivory") palms have some pertinent tracheary element details. In Ammandra roots, perforation plates are simple, 230-270 [micro]m in diameter, but in stems, only imperforate tracheary elements, 50-90 [micro]m in diameter are reported (Klotz, 1977). In Phytelephas. vessels of roots have simple perforation plates 170220 [micro]m in diameter. Klotz (1977) reports stem vessels with scalariform perforation plates in late metaxylem, 50-100 [micro]m in diameter in one collection of Phytelephas, but only imperforate tracheary elements, 30-70 gm in diameter in another. Although the difference may seem slight, the occurrence of wider late metaxylem elements in the collection observed to have vessels is suggestive. Phytelephas and Ammandra are palms of wet tropical understory habitats (Dransfield et al., 2008).

Chamaedorea is an especially interesting palm genus where analysis of habit and habitat and their relationship (or nonrelationship) to molecular trees are involved. Both Cheadle (1942) and Klotz (1977) placed Chamaedorea as a "primitive" genus in palms because it has scalariform perforation plates in vessels both in roots and stems. The root late metaxylem vessels mostly range from 30-80 [micro]m in diameter, whereas those of stems are a little wider (50-100 [micro]m). Baker et al. (2009) give Chamaedorea a "crown group" rather than an early-diverging placement within palms. Chamaedorea is usually a palm of wet understory habitats, with some species relatively small in stature. These habit/habitat features explain why late metaxylem of Chamaedorea should have narrower vessels (lower and steadier transpiration and conduction rates). If narrower vessels are more likely to have scalariform perforation plates, one can see why Chamaedorea was thought by workers who were using conceptions prior to those generated by molecular phylogenies to show more ancestral features. The role of habit and habitat as well as vessel diameter cannot be neglected in assessing the significance of perforation plate type. Chamaedorea may have had an unbroken history of occupancy of mesic habitats.

As noted above, palms do not have monocot cambia and cannot produce secondary bundles. Therefore, one would hypothesize that palms have sieve-mbe elements with great longevity, and that has, in fact, been demonstrated (Parthasarathy & Tomlinson, 1967; Parthasarathy, 1980). It has also been demonstrated in Kingia of the Dasypogonaceae, (Lamont, 1980) which produces very few if any secondary bundles. If phloem in palms has such great longevity, the xylem should also have great longevity. Davis (1961) showed that root pressure in palms could exceed 10 m, thus accounting for conductive characteristics of palms--perhaps most notably, how embolisms could be reduced and cleared should they form.

The data of Klotz (1977) reveal a strong dimorphism between early metaxylem tracheary elements of palms (frequently imperforate) and late metaxylem tracheary elements (usually vessels, much wider than early metaxylem elements). This dimorphism is suggested in transectional views (e.g., Tomlinson, 1961). This dimorphism is an ideal way of combining the conductive safety that tracheids have with the great conductive effectiveness of metaxylem vessels. Tracheids have end walls that bear pit membranes which do not permit passage of air bubbles from one tracheid to the next in a vertical series. Tyree and Zimmermann (2002) figure parenchyma surrounding large metaxylem palm vessels, a configuration that may support conduction by regulating osmotic pressure of xylem through release of photosynthates into the conductive stream, much like the concept of Sauter et al. (1973) in woody angiosperms.

Some palm leaves thought to have long scalariform perforation plates in petioles may actually have tracheids. SEM images of petiole tracheary elements of Actinokentia and Hyphorbe reveal that intact, though highly porose, pit membranes occur in the end walls of petiole tracheary elements, which are therefore arguably tracheids (Klotz, 1977).

Combined with these conductive features, palms offer mechanical division of labor that lends itself to a variety of structural types--most notably the erect unbranched stem. Vessels and tracheids in palms do not have appreciable mechanical strength compared with the bundle fibers. Fibers in bundles may be various in quantity and pattern of distribution. Palms also demonstrate various degrees of wall thickness and lignification of ground tissue, thereby also adding to mechanical strength and stem hardness (Tomlinson, 1961, 1990).

Acorus: Tracheary Elements in a Pivotal Genus

Molecular-based phylogenetic reconstructions of the monocots (e.g., Fig. 15, Davis et al., 2004) agree in placing Acorus, sole genus of Acoraceae, as the sister to the remaining monocots. This basal position makes Acorus of special interest. Does Acorus represent a xylem configuration symplesiomorphic within monocots? Does it show xylem adapted to a particular habitat? The answers to these questions are extraordinarily important because of the precedent they may set for interpretation of other monocots.

Carlquist and Schneider (1997) examined Acorus gramineus Soland. roots and found that pit membranes in end walls of tracheids have pores of various sizes. The method used, that of cutting paraffin longisections of roots, and then removing paraffin so that sections could be examined under SEM, has limitations, however. Paraffin sections are relatively thin, so that only a few fragments of end walls are available for study, and the context of these fragments is not always clear. A method that we used subsequently (Carlquist & Schneider, 201 Oa), is much better. It involves cutting longisections of liquid-preserved roots (or stems, etc.) with a single-edged razor blade, followed by changes of distilled water and drying between slides under pressure on a warming table. That method (used in original work in the present paper) has the advantage of giving thick sections, in which large portions of vessels (sometimes entire end walls) are present. Because of the thickness of the sections, one can view intact walls that have not been scraped by the blade (as well as some surfaces that have been). The thickness also results in fewer fractures (as compared to microtome sections) in pit membranes.

Using this new method, two species of Acorus are shown to have porose pit membranes in root tracheary elements (Fig. 14a-c, e-f), and material of A. calamus L. stems (Fig. 14d) had similar pit membranes. The pit membranes shown in Fig. 14 are somewhat different from those figured earlier (Carlquist & Schneider, 1997) in having pit membranes that are more network-like than porose, but the effect is very similar in both instances. Pit membranes are consistently present in end walls of tracheary elements of Acorus roots and stems. Although earlier (Schneider & Carlquist, 1998), the porousness of these pit membranes was emphasized, leading to the possibility that vessels could be said to be present, subsequent experience with SEM studies of monocots suggests that pit membrane presence of any kind in tracheary end walls is better considered under the rubric of "tracheid" or possibly "pre-vessel." Physiologically, the presence of porous or meshwork-like pit membranes would likely confine air bubbles to a single tracheary element, and this has been considered in the past to be a significant characteristic of tracheids as opposed to vessel elements.

The end walls in pit membranes in Acorus tracheids (Fig. 14) are paralleled by the conditions observed in Typha (Fig. 6), which occupies a very similar habitat. The end-wall pits in Typha roots (Fig. 6 Both genera typically grow with submersed stems and roots, but stems can also extend onto non-submerged ground provided that the soil remains moist. Acorus leaf tips turn brown readily if soil moisture falls below saturation levels. The presence of porose pit membranes in stem tracheary elements of both genera could correlate with an active conductive stream, in turn related to the sunny localities in which both grow combined with their access to abundant soil moisture.

The lateral walls of Acorus tracheary elements have pits with non-porose pit membranes (Fig. 14c), so there is differentiation between lateral walls and end walls in this respect. There are differences between end walls and lateral walls with respect to secondary wall structure: the bars between pits are slightly thinner on end walls than on lateral walls, and the end wall pits slightly wider than those of lateral walls (Fig. 14c), but the differences are not great. Such differences can be seen in tracheids of Amborellaceae and Winteraceae also.

If we are to entertain the idea that Acorus xylem represents a symplesiomorphic condition for monocots, we are forced to think that the habitat of Acorus is that of ancestral monocots, and that there has been no change over time in the ecological adaptations of Acoraceae. On the other hand, we may take a more inclusive view to the effect that Acorus, with "pre-vessel" tracheids, may be considered vesselless but with xylem expressions that correspond to occupancy of the aquatic--terrestrial interface that characterizes Acorus habitats. This more inclusive view permits us to consider monocots ancestrally vesselless, but recognizes some possibilities of diversity within the vesselless condition. Vessellessness is thus not a uniform condition, but one that is subject to variations, if not to the extent that the vessel-bearing condition is. Such an interpretation permits us to retain the concept that xylem does correspond to ecology (taking into account other water economy adaptations such as leaf shape, leaf surface moisture gradients, leaf anatomy, photosynthetic pathways, etc.).

New Paradigms

Molecular phylogenies are now guidelines for how to interpret xylem evolution. When data from habit, ecology, and physiology, xylem ontogeny, quantitative xylem features, and SEM studies are incorporated, we reach quite new perspectives from those reached by earlier workers. These should not be regarded as trends and certainly not as dicta. Each example examined in the light of such a synthesis can help us in assembling a larger picture and in exploring patterns.

What Controls Monocot Xylem Evolution?

Monocot xylem has evolved with response to ecology, but there is a wide range of ways in which clades have shown adaptation. There is not an inexorable progression of xylem specialization, as Cheadle's work (1942 et seq.) might seem to infer. For example, most bulbs have only tracheids in leaves and stems, but vessels with simple perforation plates in roots: an abrupt disjunction. Short duration of roots, persistence of succulent leaf bases, drought deciduousness of leaf blades (sometimes with attenuated longevity of leaf blades by means of succulence), and other features of this growth form must be taken into account. To say that the xylem of leaves and stems of bulbs remains "primitive" whereas that of the roots has "accelerated specialization" hides many important facts at best and may be misleading.

Although claimed to show irreversible progressions, evolution in monocot xylem can feature apparent reversions. For example, the lack of vessels in submersed aquatics such as Aponogetonaceae and Zosteraceae is an apomorphic condition, not a primitive one, and probably involves an ontogenetic foreshortening, by forming no late metaxylem.

Multiplicity of Expressions and Heterochrony as Factors in "Primitive" Xylem

Symplesiomorphic monocot xylem is a "porous" concept in that there is no single xylem formulation one can cite in contemporary monocots that represents exactly what the conductive tissue of the earliest monocotyledons was like. Monocots with numerous symplesiomorphic features in xylem have probably had long and perhaps unbroken histories of occupancy of mesic sites, but there are various ways in which mesic sites can be exploited. Modification of foliage is certainly a persistent theme, and one must take into account the multiplicity of linear/lanceolate leaf forms and their modifications. One can trace changes in DNA sequences leading away from symplesiomorphic monocots, but ancient DNA sequences do not necessarily correlate with ancient xylem structure. The global tree of monocots does permit us, however, to identify key groups that will show us some important xylem transformations. For example, the calamoid palms, identified as an early branch of the palm clade in molecular trees, have xylem adapted to particular niches, such as lianoid growth forms, that relate to major reconfiguration of their xylem. Lianas as a whole have wide stem vessels and the wider the vessel, the more the likelihood of a simple perforation plate. The lianoid calamoids have vessels that would not have been identified as "primitive" by earlier workers. Chamaedoroid palms are a crown group, but narrowness of vessels and a mesic understory habitat have favored retention in that group of xylem with long scalariform perforations that in the past have been interpreted as unqualified indicators of primitiveness. Phytelephoid palms have loss of late metaxylem pattems and therefore can have a vesselless xylem that looks more symplesiomorphic than it really is. Palms have tracheids in early metaxylem, but prominent vessels in late metaxylem. By foreshortening this developmental sequence, a kind of juvenilism, the phytelephoid palms have heterochronically attained this condition, which is compatible with their wet forest habitats.

The Paradox of "Specialized" Xylem in Early-Departing Branches of Clades

Ecological interations must be taken into account in analyzing xylem; an early-departing brach of a clades may have adapted to highly seasonal conditions and have apomorphic xylem (simple perforation plates in vessels) as have the apostasioid orchids, may retain ancient DNA sequences. Apostasioids (a small, relictual group compared to other orchids) may once have had species with long scalariform perforation plates in roots, species likely to be ancestral to orchids, but if so, those apostasioids have apparently disappeared. One can also envision a rapid ecological shift that coincided with departure of the apostasioid branch. The crown groups of orchids may have changed little in xylem, while speciating extensively. At any rate, the molecular tree of monocots must now be a guideline as to what changes in monocot xylem took place. Apostasioids are sister to the remainder of orchids (Kocyan et al., 2004). Cheadle and Tucker (1961) claimed that the apostasioids could not possibly be ancestral to other orchids, based on xylem. The possibility that living members of an early-departing clade might not represent all of the expressions that once existed in that clade should have been considered by them. As a generalization, Cheadle and Tucker (1961) claimed that "Because of its unidirectional course, vessel specialization can be used primarily in negations; that is, a taxon with a certain level (or levels) of vessel specialization cannot have been involved in the origin of another with less specialized xylem in the same organs." Other examples that may seem paradoxical until one takes a broader viewpoint include Boryaceae, Velloziaceae, and Rhipogonaceae, as noted above. There is no reason why an early-departing branch of a clade should not be, in fact, ecologically more exploitive and the remainder of the clade rather conservative in preferences.

SEM has Changed Our Ideas About Monocot Tracheary Elements

Microstructure of the end walls of tracheary elements reveals pit membranes in end walls of what appear, on the basis of light microscopy, to be vessel elements. These SEM studies show that tracheids are more widespread in monocots than the listings of Cheadle (1942); Fahn (1954) and Wagner (1977) would indicate. Among examples are a number of orchids (Carlquist & Schneider, 2006) as well as Acoraceae, Cyclanthaceae, Pandanaceae, and Philesiaceae, SEM studies of which appear in the present paper. Klotz (1977) illustrates presence of pit membranes in scalariform end walls in some palms previously thought to have scalariform perforation plates (Actinokentia, Hyphorbe). Pit membranes in end walls of tracheary elements may have various degrees of porousness, from being an open strand system to being laminar with few holes. There is no way, other than an arbitrary (and therefore probably meaningless) one, to place a boundary across this continuum.

Reports of vessel elements that have, say, 30 or more bars on the basis of light microscopy need re-examination to see whether pit membranes may be retained in the alleged perforations. End walls of longer tracheary elements are often poorly defined and may grade into lateral wall pitting, contrary to the broad-lined ovals designating end walls in the drawings in the Cheadle and Kosakai papers.

The construction of data matrices for phylogenetic purposes and other uses for anatomical data all too often demand categorization of tracheary elements as either "tracheids" or vessel elements." Vessel elements with fewer than 30 bars on end walls may be confidently recognized on the basis of light microscopy. The development of SEM in relation to tracheary element type assignment has, however, made the definitions very difficult. This is definitely a gain for understanding of the physiological nature of the conductive system. The presence of pit membranes and of pit membrane remnants in end walls makes understanding xylem evolution by relating it to ecological factors a more insightful process, because the intermediate conditions are "non-missing links." The search for definitions should not drive or distort our understanding of the evolution of structure.

Tracheary Element Length: Not an Indicator of Xylem Phylogeny in Monocots

Vessel element length and tracheid length are not phylogenetic indicators in monocots. The idea that longer tracheids and vessel elements are "more primitive" in monocots has been repeatedly promulgated (Cheadle, 1942; Cheadle & Tucker, 1961), but in fact, neither Cheadle nor any other worker really attempted to demonstrate this. Very likely, Cheadle and others found that tracheary element length is not really correlated with phylogenetic position or degree of departure from an ancestral type in monocots, and simply did not present data on this feature for that reason. Essentially, Cheadle based all of his interpretations of degree of phyletic advancement in monocots on one feature, the number of bars on a scalariform perforation plate, and, as a corollary, the distribution of various types of perforation plates systematically and organographically.

The idea that long tracheary elements are symplesiomorphic derives from the work of Bailey and Tupper (1918), who did not, in fact, study monocots at all. In woody angiosperms, Bailey and Tupper (1918) did obtain statistical correlations, but such correlations cannot be carried over to monocots. They are inapplicable in monocots because cambium is absent. The factors governing the length of fusiform cambial initials in plants with vascular cambia and the factors governing length of procambium in monocots, which lack vascular cambia, are quite different. Degree of elongation of tracheary elements in monocots is probably related to factors such as internode length, rapidity of growth, organ size, and other factors. All of these, in turn, undoubtedly have hormonal components. Research on what controls tracheary element length in monocots is needed, and can very likely be obtained by quite simple experiments and observations.

Tracheary Element End Wall Pit Membranes: Diversity, Ontogeny, and Functional Significance

Pit membranes in end walls of tracheary elements, as SEM studies show, are not all alike. If they were, we could make a clear distinction between vessel elements and tracheids, and such a distinction would probably be discernible even with light microscopy (perforation plates would be more clearly delimited than end walls of tracheids).

In fact, many tracheary elements of monocots, such as those in Orchidaceae, show all degrees of transition between laminar and non-porose pit membranes versus and presence of only a few threads as pit membrane remnants in pits of tracheary elements of end walls (Carlquist & Schneider, 2006). In genera such as Cymbidium and Phalaenopsis, pit membranes on tracheary element end walls of roots have a more open end-wall meshwork, whereas those of stems show a smaller area devoted to porosities. These is no clear delimitation between tracheids and vessel elements in such genera.

The development of vessel elements features a hydrolysis of pit membranes in the end wall (Butterfield & Meylan, 1982). Some end walls have more extensive microfibrillar networks in pit membranes than others, so when hydrolysis of the end walls occurs, some become more porose than others. A cellulosic network is not evident in the pit membranes of the end walls of maturing grass vessels (Carlquist & Schneider, 2011), but amorphous material is abundant. Microfibrillar networks in vessel element perforations, when swept away by the conductive stream, may be present in lateral ends of the perforations, probably because the center of the conductive stream exerts more pressure, whereas the microfibrils at the lateral ends are under less tension from the pressure of the stream. Different degrees of end wall hydrolysis and loss of the microfibrillar network are characteristic of different species and, in monocots, the various organs of the species. This has been demonstrated in woody angiosperms, in which even perforation plates with relatively few bars/perforations (e.g., Bruniaceae) may retain extensive pit membrane remnants (Carlquist 1992a, b).

"Neotracheids" may be a genuine phenomenon in some monocots: the tendency for pit membranes to remain intact or nearly so in pit membranes of what on the basis of having well-defined perforation plates might appear to be vessel elements. This was figured above for Ophiopogon (Fig. 5e-f) and in Orchidaceae for stems of Vanilla (Carlquist & Schneider, 2006). This phenomenon has been found in woody angiosperms such as Myrothamnaceae (Carlquist, 1988) and cannot be ascribed to lack of maturity of vessels, since it can be found repeatedly in particular species. This may be a mechanism for phylogenetic return to the conductive safety of tracheid end walls. Delimiting this phenomenon precisely from instances of pit membrane preservation in end walls of tracheary elements may not always be easy, but if pit membranes are present in apparent perforation plates that are well differentiated from lateral wall pitting, one could say that neotracheids may be present.

Adventitious Roots and the Root-Stem Juncture as a Valve

With regard to conduction in Agave deserti Engel., Ewers et al. (1992) say:

"During soil drying, the hydraulic conductance per unit pressure ([K.sub.h]) declined dramatically in the [root-stem] junctions and to a lesser extent in the roots, but not in the stems. The decline in junction [K.sub.h] was particularly important for A. deserti, which lacks vessels in its stems, because even under wet conditions, its [K.sub.h] was lower in stems and junctions than in roots....The decline in [K.sub.h] was due to embolism in the connective tracheary elements at the junction. Such connective elements may be particularly vulnerable ro embolism due to the large amount of unlignified primary cell wall. Because the embolism is reversible, the junctions act as rectifiers. The high [K.sub.h] under wet conditions allows for rapid water uptake following rainfall, and low [K.sub.h] during drought helps limit water loss from the succulent shoots to a dry soil."

Vessels cannot be continuous from adventitious roots into stems, because of their discontinuous ontogeny. The idea by Ewers et al. (1992) that the juncture between root tracheary elements and stem tracheary elements is a valve that can insure one-way flow is logical and intriguing. Agave roots have much greater duration than those of bulbs, in which functionality of roots is compressed between winter cold and summer drought. Agave is, in fact, characteristic of summer-rainfall areas in which winters are dry but without severe frost. Roots and leaves with long duration thus are correlated with the advantages and limitations of the xylem in Agavaceae. The ephemeral nature of roots in bulbs may be more important than the valve function with respect to their short growth and flowering season.

The Ewers et al. (1992) concepts do explain why monocots have been able to occupy such a wide range of habitats despite the limitations of adventitious roots. The nature of adventitious roots requires them to draw from shallower soil levels than can deep taproots in woody angiosperms. The idea that roots could lose water to dry soil is probably not a function of the nature of xylem, but based on other facts.

There is broader significance in the Ewers et al. (1992) paper in that it joins data from physiology with information about xylem anatomy. Most of our data on conductive physiology are based on woody angiosperms, and the applicability of physiological data from woody growth forms with high transpiration rates to monocots needs to be considered.

Vessellessness in Monocots: Conductive Safety, Structural Necessity, or Relictualism?

Cheadle (1942, 1943a, b) envisioned the upward evolutionary progression of vessels within the organs of monocotyledons: first roots, then stems, then inflorescence axes, and then leaves. Such progression is not as inexorable as this scheme might suggest, and Cheadle himself (1942) calls attention to some genera in which the progression is not so simple (e.g., Dracaena).

Much attention has been focused on vesselless woody angiosperms, but little attention has been paid to why vessellessness is so pervasive in monocots. The safety of the all-tracheid system and the physiological limitations that early vessel-bearing angiosperms may have encountered have been noted earlier (Carlquist, 1975, 1984a, 1985, 1988, 2012; Hacke et al., 2007; Sperry et al., 2007). There is a trade-off in the ability of tracheids to resist formation and cell-to-cell transmission of embolisms on the one hand and the inability of tracheids to be ideal capillaries that can deal with peak flow and conduction speed on the other. Succulence, boundary layers (epidermis, cuticle) highly resistant to high transpiration rates, narrowness of leaves, crassulacean acid metabolism, and C4 photosynthesis, as noted earlier, are all mechanisms that are in play in the monocots with all-tracheid systems in stems and leaves. We need to study xylem anatomy in relation to transpirational and flow characteristics.

In this regard, one notes that monocots with vessels in stems and leaves may have some characteristics different from those of monocots that have only tracheids in stems and leaves. Palms are an interesting example of a growth form with ecological and habital capabilities that are different from those of the monocots with vesselless stems and leaves. They, like grasses, may have conductive systems that operate to an appreciable extent on root pressure, and thereby are able to occupy habitats different from those of the monocots without vessels in stems and leaves.

Arborescence: Xylary Strategies and Limitations in Monocots

There are two major kinds of arborescence within monocots: the palm habit (with absence of secondary bundles); and other arborescent forms with secondary bundles. Palms have wide late metaxylem vessel elements in roots and stems. Such vessels can be correlated with high transpirational demands, which in turn are related to the massive size of leaves (which in turn are correlated with infrequency of branching in palms). Lack of secondary bundles in palms requires compensation in the form of great longevity in palm xylem and phloem, which can have durations that exceed 100 years. Palms are capable of unusually high root pressures, which may be achieved by with the aid of parenchyma sheaths that can mediate ion and photosynthate content of vessels. The bundles in palm stems contain a xylary dimorphism: imperforate early metaxylem tracheary elements (= tracheids) combined with wide late metaxylem vessels. This dimorphism provides both conductive safety and maximal flow rates. The syndrome of palm stems and roots and their xylem features does not occur in other monocots, although Kingia (Dasypogonaceae) and Pandanus (Pandanaceae) have some of the same features.

Other arborescent monocots have vesselless stems and leaves, mostly narrow leaves (some species of Cordyline excepted), that are usually slender and have various mechanisms for restricting transpiration (Napp-Zinn, 1984). In addition to various degrees of succulence (notably in Aloe), one sees thick cuticles (Yucca), and stomata sunken in grooves (Beaucarnea, Hesperoyucca) in the non-palm tree monocots. The transpiration limitations of such leaves correlate with the all-tracheid constitution of the stem and leaf vascular bundles. Branching is much more frequent in non-palm monocots than in palms.

The presence of a lateral meristem, here called the monocot cambium, which produces secondary bundles in the non-palm monocots, adds to the conductive capabilities of stems, and in most of these monocots, by far the majority of bundles are secondary ones. The xylem in such bundles, however, consists wholly of tracheids. This is a consequence of the adventitious habit of roots, because there is no way for vessels in adventitious roots to connect with those of stems. Secondary bundles are also formed on roots in Dracaena draco and other species of Dracaena.

The non-palm arborescent monocots thus have limitations in plant size, flow fluctuation capabilities, transpiration, restriction in number of leaves, and leaf surface area. These limitations are balanced against the capabilities of an all-tracheid stem and leaf xylem to resist spread of air embolisms from one xylem cell to another, and the all-tracheid formula is in turn necessitated by the fact that vessels in roots must end blindly where they contact stem bundles because of the adventitious nature of roots. One should also take into account other features that factor into this formula, such as crassulacean acid metabolism, which has been shown to exist in Agave, Aloe. and Yucca (Nelson et al., 2005).

The Scalariform System in Angiosperm Tracheary Elements and Its Significance

While celebrating the remarkable success of the conifer torus-margo system in coniferous tracheid pits, Pitterman et al. (2005) overlook the underlying significance of associated pit shape. The torus-margo system works in conifers (including Gnetales) because the pits are circular. Thus, the slender threads of the margo (which offer maximal conductive space between them for the margo holes) can carry the central toms easily, because stress is distributed evenly throughout the margo. Displacement of the torus to the pit aperture is thus easily accomplished. The great length of the end wall of a conifer tracheid expands the conductive capabilities of the coniferous pit very considerably. Radial widening of a conifer tracheid is not accompanied by widening of pits. Instead, it is accompanied by replication of circular pits, so that two or three series of circular bordered pits may be observed on some conifer walls (Bailey, 1925), such as those of Sequoia. Pit redundancy rather than pit widening is the strategy seen in the pits of coniferous tracheids (and in the vessels of Ephedra). Even in primary xylem of conifers and Gnetales, one sees circular bordered pits intercalated into the helical bands

Vesselless woody angiosperms all have scalariform widening of pits on end walls of wider tracheids: Amborella (Carlquist & Schneider, 2001), Trochodendraceae (including Tetracentron: Carlquist, 1988), and some Winteraceae (notably Tasmannia and Zygogynum, but not Exospermum or Pseudowintera: Carlquist, 1989). All of the early-diverging vessel-bearing woody angiosperms have scalariform perforation plates on end walls of vessels, and this is certainly true of vessels in early-diverging monocots as well.

Angiosperms seem, from their point of origin onwards, not to have had the genetic information or developmental pathways to replicate the conifer pattern. From the outset, the conifer-margo pattern seen in conifers was absent.

The formation of scalariform pits in end walls of vessels and tracheids in angiosperms does not permit a torus-margo system like that of conifers. The stresses on margo threads would not be equally distributed on the margo, and thus the toms would not be a reliable closure device. Instead, angiosperms have invented pit membranes of various thicknesses. While some of these have what have been called tori or pseudotori (Dute & Rushing, 1987; Rabaey et al. 2006), they lack the important feature of a coniferous torus-margo system: the wide margo spaces that enhance conduction. One compensation for this lack, in vesselless species, is thinness of pit membranes. The pit membranes of Amborella are very thin and break easily during handling, so some have even considered them to be absent at maturity (Feild et al., 2000), but in fact, they are present if handled correctly and are highly porous (Carlquist & Schneider, 2001). This is true in Bubbia also (Carlquist, 1983). Thin, easily broken but porous pit membranes may be tolerable because in the habitats where these vesselless angiosperms grow, deflection of the pit membranes (due to pressure changes among tracheids) is likely to be minimal. Thin porous membranes in end walls of tracheids--and then vessels--may be the angiospermous compensation for lack of the margo spaces that conifers have. There is no differentiation between end walls and lateral walls in coniferous tracheids. Jansen et al. (2009) have shown that thinner pit membranes (as on vessel element end walls) have larger pores. In angiosperms, however, lateral walls of vessels have relatively non-porous (and variously thick) pit membranes, which resist deflection and aspiration at the cost of lowered conduction of water. Conduction is achieved by the porous end walls. This is shown well by orchids (Carlquist & Schneider, 2006; see above also). The selective pressure to develop wider tracheids (and thereby scalariform end walls, and subsequent to that, scalariform perforation plates in vessel elements) in angiosperms comes from the value of wider capillaries (conductivity equal to the fourth power of the diameter, the Hagen-Poiseuille equation), so that widened capillaries of vessels, even if only a little wider than tracheids, have enhanced conductivity by virtue of the total area of the perforations on the end walls, and the permeability to water flow of the thin porous pit membranes in the end walls. Ellerby and Ennos (1998) find that resistance of the end walls is not substantial conpared to the resistivity to conduction of the lateral wails in angiosperm vessels. Hacke et al. (2007) offer a dimmer view of the advantage of the vessel, at least in its early evolutionary states. However, more primitive woods tend to occur in such places as cloud forests with wet soil and humid air, where demands on the conductive system for conductive efficiency are modest.

Thus angiosperms have exploited a scalariform system with advantages and disadvantages quite unlike those of the circular bordered pit system of gymnosperms. Similarities in the tori or pseudotori of some angiosperm pits (Dute & Rushing, 1987; Rabaey et al., 2006) to the coniferous pit, with its highly conductive margo portion, are misleading. Angioperm toil and pseudotori may aid in pit aspiration (closure), but the accompanying margos are not at all like those of conifers in conductive ability, and the great area of pits on conifer end walls has no counterpart in angiosperms, with tori or pseudotori. Angiosperms have followed paths of end wall dimorphism (perforation plates different from lateral wall pitting) and varied thicknesses of pit membranes, a series of adaptations entirely different from those of conifers.

Conductive Problems in Monocots: What is Involved?

The conductive system of vascular plants consists of dead cells--tracheids and/or vessels. Or does it? Experimental work on how xylem functions has been done mostly on woody species, because they are more convenient experimental material. In woody angiosperms, an emerging picture trends to show that early angiosperms have high resistivity and low vulnerability in xylem: they conduct less efficiently but cavitate infrequently (Carlquist, 2012). This Vend is not without exceptions--and ecological shifts that may be accompanied by new structural modes. In clades that have moved into drier habitats, the balance frequently shifts toward low resistivity but with higher vulnerability--daily cavitations occur and must be refilled (Vogt, 2001). This seeming paradox is explained by the fact that in order to capture one molecule of C[O.sup.2], a plant may have to lose several hundred water molecules (Jones, 1992), a process colorfully described by Holbrook et al. (2002) as "foraging" with an attendant cost. Risk reduction can occur in the form of CAM photosynthesis (Agave) or C4 photosynthesis (grasses), in which nocturnal opening of stomata involves lowered loss of water while achieving C[O.sup.2] absorption.. Cavitation can be repaired even while a plant experiences negative pressures in xylem, as demonstrated in Oryza (Stiller et al., 2005).

Risk reduction is not risk elimination. How do plants reverse embolisms if they occur frequently, even daily? In the case of woody plants, the answer appears to reside in the activity of axial parenchyma cells that accompany vessels (Holbrook et al., 2002). Indeed, there is no other obvious explanation for the common occurrence of axial parenchyma in woody angiosperms (Carlquist, 2012)--axial parenchyma can serve in water or photosynthate storage, but that is not prominent in most woods, nor would it explain the vessel-centered distribution of most axial parenchyma in woods that do not have a tracheid background.. In the case of monocots, axial parenchyma is not present, but metaxylem vessels are often sheathed with parenchyma (Metcalfe, 1960; Tomlinson, 1961, 1969, 1990) that may have an equivalent function.

As stated by van Ieperen (2007), "there is increasing support for the idea that ions in xylem sap can influence the hydraulic conductance of the xylem in plants." Porous pit membranes may permit water flow from one tracheary element to another, but at a cost: more than 80 % of the resistance to flow in xylem is provided by the pit membranes (Tyree & Zimmermann, 2002; Choat et al., 2006). Is there any way of changing this in a living plant? One recent idea (Holbrook et al., 2002) cites "hydrogels.": "With increasing concentrations of ions, these hydrogels are hypothesized to shrink, increasing the porosity of the pit membrane and thus decreasing its resistance to water flow" (Holbrook et al., 2002). The implications of this for the pores in end walls of monocot tracheids are especially interesting, because of the abundance of vesselless tissues in aerial portions in so many monocots. Transfer of water from one tracheid to the next in a vertical series is mediated by pores in end walls--which both confer safety (confining air bubbles to individual tracheids) and decrease flow. The implications also apply to vessels, in that perforations plates permit maximal conductive flow but lateral wall pits are often in contact with other vessels in monocots. The porous pit membranes in end walls of vessel elements in some monocots, such as those of orchids and others cited in this paper, may relate to hydrogel presence, and explain why pit membranes have been retained in end walls of otherwise vessel-like tracheary elements of some monocots.

Monocot Origins

Molecular phylogenies as well as anatomical evidence favor the origin of monocots from vessel-bearing basal angiosperms. The contention by Cheadle (1953) that vessels evolved independently in monocots and dicots can no longer be supported--at least in the way Cheadle envisioned it. All recent global molecular phylogenies of angiosperms (e.g., APG III, 2009) show monocots nested within basal angiosperms, all of which except for Amborellales and Nymphaeales are considered to have vessels (or to have lost them secondarily, as in Winteraceae). The angiosperms closest to the origin of monocots include Chloranthaceae and the Piperales (Aristolochiaceae, Lactoridaceae, Piperaceae, and Saururaceae), all of which do have vessels. By reduction of cambial activity in stems like those of Chloranthaceae or Saururaceae, one can achieve monocot-like stem structure (Carlquist, 1992a, b, 2009) with respect to bundles and ground tissue. The three-dimensional patterns of bundles in these families differ from those of monocots. Certainly pertinent is the fact that Chloranthales and Piperales are sympodial and have adventitious roots (Carlquist, 2009), circumstances probably basic to monocots, in which those interrelated features are symplesiomorphic. One could also imagine very long scalariform vessel elements in monocots, such as those of Astelia, to have been derived from similar vessels in Chloranthaceae.

Is it conceivable that the earliest monocots may have been vesselless? That possibility cannot be ruled out entirely, and Chloranthaceae show why. Primary xylem and, in the stems, even secondary xylem of Sarcandra is vesselless, as claimed by Bailey and Swamy (1950). The report of Takahashi (1988) that shows some alteration in a few pit membranes of secondary xylem stem tracheids of Sarcandra, based on TEM, does not give convincing evidence. There is no differentiation among the stem tracheids of Sarcandra, all of which are monomorphic in diameter and pitting. There is dimorphism in tracheary elements in the secondary xylem of roots of Sarcandra, and vessels are demonstrably present there (Carlquist, 1987, 2009). However, if the secondary xylem of chloranthoids like Sarcandra were reduced phylogenetically, one would obtain a vesselless condition.

Imagining why a chloranthoid such as Sarcandra would lose vessels is another matter. Saururaceae, the marsh-inhabiting close relatives of Chloranthaeeae, have nearly lost cambial activity in stems, but they have not lost metaxylem vessels. Progression to the habit of a submersed aquatic would probably be a necessary step in such a phylesis. Genera of Alismatales with emergent leaves, such as Sagittaria, have vessels in roots (Wagner, 1977; Tomlinson, 1983).

Acorus, which has generally been considered sister to the remaining monocots (Davis et al., 2004; APG III, 2009) may prove, depending on one's terminology, to be entirely vesselless or with tracheidlike vessels, because pit membranes on tracheary element end walls in roots are reticulate to somewhat porous on end walls of roots, and vessels are lacking in stems (Carlquist & Schneider, 1997); see also above. Cheadle (1942) regarded Acorus as having vessels in roots, but he did not have the benefit of SEM, which proves crucial in these matters.

The next most basal branch of the monocot tree (Fig. 14), Arales, does have vessels in roots (Cheadle, 1942, Carlquist & Schneider, 1998; Schneider & Carlquist, 1998), although the tribe Orontieae, sister to the remaining Araceae (French et al., 1995) has not been monographed with respect to xylem.

If one imagines a scenario in which monocots originated in an aquatic environment and had little metaxylem, and subsequently acquired late metaxylem with tracheary elements wide enough to develop into vessels, monocots could have originated in an aquatic environment--probably with stems mostly submersed. The problem with this "mostly submersed" scenario is that basal pre-monocot angiosperms would have had to have entered the aquatic environment, perhaps as submersed aquatics to account for vessel absence. Niches for submersed aquatics are relatively few and require a series of concurrent adaptations for dealing with low oxygenation and other features. In any case, the vast radiation of monocots would have occurred after vessels were acquired according to this scenario. Origin of vessels is conceivable in an aquatic monocot that has active transpiration in leaves but grows in moist soil. Acoraceae and Typhaceae exemplify functionally vesselless monocots that have what I would consider tracheids representing a "pre-vessel" condition. One should note that various degrees of intermediacy in expression must be reflected in terminology now that we have SEM information. We can no longer pretend that light microscopy can be the sole method on which concepts of tracheary elements is based.

The other available hypothesis would place monocot origin in moist terrestrial environments, such as at the margins of streams and lakes, or in marshes subject to water fluctuation. In this "moist terrestrial" scenario, monocots would have originated from basal angiosperms that had limited vascular cambial activity. This would explain why early monocots would have vessels in roots (an adaptive feature that permits rapid uptake of water rapidly from soil as it dries or begins to dry). Campynemataceae, Taccaceae, and Lapageria of the Philesiaceae might exemplify cases of vesselless monocots living in moist but not aquatic environments.

Vessellessness in shoots and leaves characterizes many species of monocots (chiefly Arales, Liliales, and Asparagales), and is often coupled with devices to limit transpiration. The disjunction between vessel-bearing roots and vesselless stems has certainly not deterred radiation of monocots, which have taken advantage of this feature in habitats with frequently-changing moisture availability. Vessels in adventitious roots cannot connect to vessels in stems, so adventitious roots and the associated sympodial (often rhizomatous or prostrate) branching systems associated with adventitious roots in monocots have a water conduction system in which vessellessness in stems and leaves is not disadvantageous, provided that high peak volumes of water are not required to be transferred from roots into stems. The earliest angiosperms appear to have been sympodial (Carlquist, 2009). There are many favorable habitats for sympodial growth forms, as shown by both monocots and ferns. Taproots are not adaptive in environments which become so moist that oxygenation is problematic. Prostrate sympodial growth forms are advantageous in this case. Prostrate sympodial stem systems are also excellent for colonizing wider areas, in contrast to the territorial restriction imposed by having a taproot system coupled with upright stems.

The differences between sympodial systems with adventitious roots and monopodial systems with taproots are profound. Monocots certainly exemplify the former, and show some unappreciated correlations. In a prostrate sympodial system, available moisture is limited to relatively superficial soil layers--or to aquatic and semi-aquatic systems (palms and a few other monocots are exceptions). Related to this, very likely, is the vesselless condition of many monocot stems and leaves, because a tracheidonly system offers conductive safety that can counter the fluctuations in moisture ability of upper soil levels (taproots are advantageous where a water table can be reached by roots). Restricted leaf surface (and other foliar modifications) are related to the sympodial habit by virtue of greater fluctuation in moisture availability also: the linear to lanceolate shape of leaves of many monocots exemplifies this, but such leaves, as with conifer needles, are adaptive where light levels are high. Broader, netveined leaves in monocots (most commonly found in species of shady habitats) have been shown to have originated at least 26 times and lost eight times in monocots (Givnish et al., 2005).

The lack of cambium in monocots can be understood in terms of the progressive death and decay of older portions of the sympodial system. Investment in secondary tissue would not be economical if any given rhizome portion is relatively short-lived, and newer portions are served by adventitious roots. To be sure, monocots that produce additional bundles from a monocot cambium are exceptions to this, but such monocots can be seen as having superimposed a new vascular formula onto a basic monocot vascular system as seen in, say, Arales or Zingiberales. Monocots with additional bundles are also uptight, and therefore tend to escape from the limitations of the prostrate system with respect to light-gathering capabilities and also, indirectly, water acquisition (roots can have longer duration if attached to stems of indefinite longevity). These considerations, as well as the molecular evidence, show that arborescence does not represent the ancestral life form in monocots. The idea that a monocot cambium ("secondary thickening meristem') was a symplesiomorphic feature in monocots was proposed by Rudall (1991), but is not supported by recent phylogenetic work.

The differences between the "aquatic" and "moist terrestrial" hypotheses are not great. The differences between tracheids with porous membranes in end walls and vessel elements in which pit membranes are swept away by the conductive stream likewise are not great. The numerous shifts and radiations in early basal angiosperms have probably masked forever the precise nature of the earliest angiosperms in general, and the earliest monocots in particular. We may wish for plants to show us ancient structural modes, but in fact, survival to the present demands constant modifications that meet contemporary requirements. Comparative anatomy of living plants should be interpreted primarily in terms of present-day functions and ecology. Plant structure does not evolve independently of function, ecology, and habit. The case of Acorus shows us that vesselless and vessel-beating conditions are not sharply defined (a fact made abundantly clear by SEM studies), but both offer a range of possible structures. Trying to define vesselless and vessel-bearing so as to be mutually exclusive not only runs counter to observed fact, it makes the evolutionary shifts in xylem evolution more difficult to understand.

Acknowledgements The encouragement and help of Edward L. Schneider, with whom I collaborated earlier in SEM studies of monocot xylem, are much appreciated. Materials for original SEM data reported here came from various sources, notably including the National Tropical Botanical Garden, the Fairchild Tropical Garden, and the Lotusland Foundation. David Lorence, Ken Wood, Kevin Kenneally, and Virginia Hayes deserve special thanks for collecting and sending liquid-preserved materials. Thomas J. Givnish kindly gave permission for use of the phylogenetic tree reproduced in Fig. 15. Dr. Larry H. Klotz, of Shippensburg University, gave me a copy of his 1977 Cornell thesis on palm xylem, and this unpublished document, rich in data, has been of substantial importance. John Garvey deserves special acknowledgement for his aid in preparation of illustrations. Dennis W. Stevenson provided valuable editorial support.

Literature Cited

APG III. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161: 105-121.

Bailey, I. W. 1925. Some salient lines of specialization in tracheary pitting. I. Gymnospermae. Annals of Botany 39: 587-598.

--1944. The development of vessels in angiosperms in morphological research. American Journal of Botany 31:421-428.

--& B. G. L. Swamy. 1950. Sarcandra, a vesselless genus of the Chloranthaceae. Journal of the Arnold Arboretum 31:117-129.

--& W. W. Tupper. 1918. Size variation in tracheary cells. I. A comparison between the secondary xylems of vascular cryptogams, gymnosperms, and angiosperms. Proceedings of the American Academy of Arts and Sciences 54: 149-204.

Baker, W. J., V. Savolainen, C. B. Asmussen-Lange, M. W. Chase, J. Dransfield, F. Forest, M. M. Harley, N. W. Uhl & M. Wilkinson. 2009. Complete generic level phylogenetic analyses of palms (Arecaceae) with comparisons of ssupertree and supermatrix approaches. Systematic Botany 58: 240-256.

Bessey, C. E. 1915. The phylogenetic taxonomy of flowering plants. Annals of the Missouri Botanical Garden 2:109-164.

Bierhorst, D. W. & P. M. Zamora. 1965. Primary xylem elements and element associations of angiosperms. American Journal of Botany 52:657-710.

Butterfield, B. G. & B. A. Meylan. 1982. Cell wall hydrolysis in the tracheary elements of the secondary xylem. In: P. Baas (ed). New perspectives in wood anatomy (pp 71-84). Martinus Nijhoff, Publishers, The Hague.

Carlquist, S. 1975. Ecological strategies of xylem evolution. University of California Press, Berkeley.

--1980. Further concepts in ecological wood anatomy, with comments on recent work in wood anatomy and evolution. Aliso 9: 499-553.

--1983. Wood anatomy of Bubbia, with comments on the origin of vessels in dicotyledons. American Journal of Botany 70: 578-590.

--1984a. Vessel grouping in dicotyledons: significance and relationship to imperforate tracheary elements. Aliso 10: 505-525.

--1984b. Wood anatomy of Lardizabalaceae, with comments on the vining habit, ecology, and systematics. Botanical Journal of the Linnean Society 88: 257-277.

--1985. Vasicentric tracheids as a drought survival mechanism in the woody flora of southern California and similar regions; review of vasicentric tracheids. Aliso 11:37-68.

--1987. Presence of vessels in Sarcandra (Chloranthaceae); comments on vessel origins in angiosperms. American Journal of Botany 74: 1765-1771.

--1988. Comparative wood anatomy. Springer Verlag, Heidelberg.

--1989. Wood anatomy of Tasmannia; summary of wood anatomy of Winteraceae. Aliso 12: 257-275.

--1992a. Pit membrane remnants in perforation plates of primitive dicotyledons and their significance. American Journal of Botany 79: 660-672.

--1992b. Wood anatomy of Chloranthus; summary of wood anatomy of Chloranthaceae, with comments on vessellessness, and the origin of monocotyledons. IAWA Bulletin, new series 13:3-16.

--2007. Successive cambia revisited: ontogeny, histology, diversity, and functional significance. Journal of the Torrey Botanical Society 134:301-332.

--2009. Xylem heterochrony: an unappreciated key to angiosperm origins and diversification. Botanical Journal of the Linnean Society 161: 26-65.

--2010. Caryophyllales: a key group for understanding wood anatomy characters and their evolution. Botanical Journal of the Linnean Society 164: 342-393.

--2012. How wood evolves: a new synthesis. Botany (in press). Carlquist, S. & E. L. Schneider. 1997. Origins and nature of vessels in monocotyledons. 1. Acorus. International Journal of Plant Sciences 158:51-56.

--&--. 1998. Origins and nature of vessels in monocotyledons. 5. Araceae subfamily Colocasioideae. Botanical Journal of the Linnean Society 128: 71-86.

--&--. 2001. Vegetative anatomy of the New Caledonian endemic Amborella trichopoda. New data, relationships with Illiciaceae, and implications for vessel origin and definition. Pacific Science 55: 305-312.

--&--. 2006. Origins and nature of vessels in monocotyledons. 8. Orchidaceae. American Journal of Botany 93: 963-971.

--&--. 2010a. Origins and nature of vessels in monocotyledons. 1. Primary xylem microstruc-ture, with examples from Zingiberales. International Journal of Plant Sciences 171:258-266.

--&--. 2010b. Origins and nature of vessels in monocotyledons. 12. Pit membrane microstructure diversity in tracheary elements of Astelia. Pacific Science 64: 607-618.

--&--. 2011. Origins and nature of vessels in monocotyledons. 13. Scanning electron microscopy studies of xylem in large grasses. International Journal of Plant Sciences 172:

Carlquist, S., E. L. Schneider & K. Kenneally. 2008. Origins and nature of vessels in monocotyledons. 10. Boryaceae: xeromorphic xylem structure in a resurrection plant. Journal of the Royal Society of Western Australia 91: 13-20.

--&S. Zona. 1988. Wood anatomy of Papveraceae, with comments on vessel restriction paterns. IAWA Bulletin new series 9: 253-267.

Chase, M. 2004. Monocot relationships: an overview. American Journal of Botany 91:1645-1655.

--, et al. 1993. Phylogenetics of seed plants: an analysis of nucleotide sequences from the plastid gene rbcL. Annals of the Missouri Botanical Garden 80: 528-580.

--, M. F. Fay, D. S. Devey, O. Maurin, N. Ronsted, J. Davies, Y. Pillon, G. Peterson, O. Seberg, M. N. Tamura, C. B. Asmussen, K. Hilu, T. Borsch, J. I. Davis, D. W. Stevenson, J. C. Pires, T. J. Givnish, K. J. Sytsma, M. M. McPherson, S. W. Graham & H. S. Rai. 2006. Multi-gene analyses of monocot relationships: a summary. Pp 63-75. In: J. T. Columbus, E. A. Friar, J. M. Porter, L. M. Prince, & M. G. Simpson (eds) Monocots: Comparative Biology and Evolution (including Poales). Rancho Santa Ana Botanic Garden, Claremont, CA.

Cheadle, V. I. 1937. Secondary growth by means of a thickening ring in certain monocotyledons. Botanical Gazette 98: 535-555.

--1942. The occurrence and types of vessels in the various organs of the plant in the Monocotyledoneae. American Journal of Botany 29: 441-450.

--1943a. The origin and certain trends of specialization of the vessel in the Monocotyledoneae. American Journal of Botany 30:11-17

--1943b. Vessel specialization in the late metaxylem of the various organs in the Monocotyledoneae. American Journal of Botany 30: 484-490.

--1953. Independent origin of vessels in the monocotyledons and dicotyledons. Phytomorphology 3: 23-44.

--1963. Vessels in Iridaceae. Phytomorphology 13: 245-248.

--1968. Vessels in Haemodorales. Phytomorphology 18: 412-420.

--& H. Kosakai. 1971. Vessels in Liliaceae. Phytomorphology 21: 320-333.

--& J. M. Tucker. 1961. Vessels and phylogeny of Monocotyledoneae. In Recent Advances in Botany (pp. 161-165). University of Toronto Press, Toronto.

Choat, B., T. VC. Brodie, A. R. Cobb, M. A. Zieniecki & N. M. Holbrook. 2006. Direct measurement of intervessel pit membrane hydraulic resistance in two angiosperm tree species. American Journal of Botany 93: 993-1000.

Clearwater, M. J. & G. Goldstein. 2005. Embolism repair and long distance water transport. Pp 375-400. In: N. M. Holbrook & M. A. Zwieniecki (eds). Vascular transport in plants. Elsevier Academic Press, Oxford.

Davis, T. A. 1961. High root pressures in palms. Nature 192: 277-278.

Davis, J. I., D. W. Stevenson, G. Petersen, O. Seberg, L. M. Campbell, J. V. Freudenstein, D. H. Goldman, C. H. Hardy, F. A. Michelangeli, M. P. Simmons, C. D. Specht, F. Vergara-Silva & M. Gandolfo. 2004. A phylogeny of the monocots, as inferred from rbcL and atpA sequence variation, and a comparison of methods for calculating jacknife and bootstrap values. Systematic Botany 29:467-510.

Dransfield, J., N. Uhl, C. Asmussen, W. J. Baker, M. Harley & C. Lewis. 2008. Genera palmarum. The evolution and clasification of palms. Royal Botanic Garden, Kew, Kew Publishing.

Dute, R. R. & A. E. Rushing. 1987. Pit pairs with tori in the wood of smanthus americanus. 1AWA Bulletin, new series 8: 237-244.

Ellerby, D. J. & A. R. Ennos. 1998. Resistance to fluid flow of model xylem vessels with simple and scalariform perforation plates. Journal of Experimental Botany 49:979-985.

Ewers, F. W. 1985. Xylem structure and water conduction in conifer trees, dicot trees, and lianas. IAWA Bulletin, new series 6: 309-317.

--, G. B. North & P. S. Nobel. 1992. Root-stem junctions of a desert monocotyledon and a dicotyledon: hydraulic consequences under wet conditions and during drought New Phytologist 121:377-385

--, H. Cochard & M. T. Tyree. 1997. A survey of root pressures in vines of a tropical lowland forest. Oecologia 110: 191-196.

Fahn, A. 1954. Metaxylem elements in some families of the Monocotyledoneae. New Phytologist 53: 530-540.

Feild, T. S., H. A. Zwieniecki, T. Brodribb, T. Jeffrey, M. J. Donoghue & N. M. Holbrook. 2000. Structure and function of tracheary elements in Amborella trichopoda International Journal of Plant Sciences 161: 705-712.

Fisher, J., G. Angeles, F. W. Ewers & J. Lopez-Portillo. 1997a. Survey of root pressure in tropical vines and woody species. International Journal of Plant Sciences. 158: 44-50.

Fisher, J. B., H. Cochard & M. T. Tyree. 1997b. A survey of root pressures in vines of a tropical lowland forest. Oecologia 110: 191-196.

French, J. C., M. G. Chung & Y. K. Hur. 1995. Chloroplast DNA phylogeny of the Ariflorae. Pp 255-175. In: P. J. Rudall, P. J. Cribb, D. F. CutLer, & C. J. Humphries (eds). Monocotyledons: systematics and evolution. Royal Botanic Gardens, Kew

Frost, F. H. 1930a. Specialization in secondary xylem in dicotyledons. I. Origin of vessel. Botanical Gazette 89: 67-94.

--1930b. Specialization in secondary xylem in dicotyledons. II. Evolution of end wall of vessel segment. Botanical Gazette 90:198-212.

--1931. Specialization in secondary xylem in dicotyledons. III. Specialization of lateral wall of vessel segment. Botanical Gazette 91: 88-96.

Givnish, T. J., J. C. Pires, S. W. Graham, M. A. McPherson, L. M. Prince, T. B. Patterson, H. S. Rai, E. H. Roalson, T. M. Evans, W. J. Hahn, K. C. Millam, A. W. Meerow, M. Molivray, P. J. Kores, H. E. O'Brien, J. C. Hall, W. J. Kress & K. J. Sytsma. 2005. Repeated evolution of net venation and fleshy fruits among monocots in shaded habitats confirms a priori predictions: evidence from an ndhF phylogeny. Procedings of the Royal Society B 272: 1481-1490.

--, J. H. Leebens-Mack, M. Ames Sevillano, J. R. McNeal, P. R. Steele, J. L Davis & C. Ane. 2010. Assembling the tree of the monocotyledons: plastome sequence phylogeny and evolution of Poales. Annals of the Missouri Botanical Garden 87: 584-616.

Hacke, U., J. S. Sperry, T. S. Feild, Y. Sano, E. H. Sikkema & J. Pitterman. 2007. Water transport in vesselless angiosperms: conductive efficiency and cavitation safety. International Journal of Plant Sciences 1168:1113-1126.

Hargrave, K. R., K. L. Kolb, F. W. Ewers & S. D. Davis. 1994. Conduit diameter and drought-induced embolism in Salvia mellifera Greene (Labiatae). New Phytologist 126: 695-705.

Holbrook, N. M. & M. A. Zwieniecki. 1999. Embolism repair and xylem tension: do we need a miracle? Plant Physiology 120: 7-10.

--,--& P. J. Melcher. 2002. The dynamics of "dead wood": maintenance of water transport through plant stems. Integrative and Comparative Biology 42: 493-496.

Jansen, S., B. Choat & A. Pletsers. 2009. Morphological variation of intervessel pit membranes and implications to xylem function in angiosperms. American Journal of Botany 96: 409-419.

Jones, H. G. 1992. Plants and microclimate: a quantitative approach to environmental plant physiology. Cambridge University Press, Cambridge.

Judd, W. S., W. L. Stern & V. I. Cheadle. 1993. Phylogentic position of Apostasia and Neuwiedia (Orchidaceae). Botanical Journal of the Linnean Society 13:87-94.

Keating, R. C. 2003. The anatomy of monocotyledons. IX. Acoraceae and Araceae. Clarendon, Oxford.

Klotz, L. H. 1977. A systematic survey of the morphology of tracheary elements in palms. Ph. D. Thesis, Cornell University, Ithaca.

Kocyan, A., Y.-L. Qin, P. K. Endress & E. Conti. 2004. A phylogenetic analysis of Apostasioideae based on ITS, trnL-F and matK sequences. Plant Sytematics and Evolution 247: 203-213.

Kohonen, M. M. & A. Helland. 2009. On the function of wall sculpturing in xylem conduits. Journal of Bionic Engineering 6: 324-329.

Kribs, D. A. 1935. Salient lines of structural specialization in the wood rays of dicotyledons. Botanical Gazette 96: 547-557.

--1937. Salient lines of structural specialization in the wood parenchyma of dicotyledons. Bulletin of the Torrey Botanical Club 64: 177-186.

Lamont, B. B. 1980. Tissue Iongevity of the arborescent monocotyledon, Kingia australis (Xanthorrhoeaceae). American Journal of Botany 67:1262-1264.

McCully, M. E., C. X. Huang & L. E. C. Ling. 1998. Daily embolism and refilling of xylem vessels in the roots of field-grown maize. New Phytologist 138: 327-342.

Metcalfe, C. R. 1960. Anatomy of the monocotyledons. I. Gramineae. Clarendon, Oxford.

Napp-Zinu, K. 1984. Anatomic des Blattes. II. Angiospermen Band 2. Experimentelle und okologische Anatomic des Angiospermblattes. Handbuch der Pflanzenanatomie. Gehruder Borntraeger, Berlin.

Nelson, E. A., T. L. Sage & R. F. Sage. 2005. Functional leaf anatomy of plants with crassulacean acid metabolism. Functional Plant Biology 32: 409-419.

Nobel, P. S. 1988. Environmental biology of agaves and cacti. University of California Press, Berkeley.

--4& T. L. Hartsock. 1978. Resistance analysis of nocturnal carbon dioxide uptake by a crassulacean acid metabolism succulent, Agave deserti. Plant Physiology 61:510-514.

Parthasarathy, M. V. 1980. Mature phloem of perennial monocotyledons. Berichte der deutschen botanischen Gesellschaft 93: 57-70.

--& P. B. Tomlinson. 1967. Anatomical features of metaphloem in stems of Sabal, Cocos and two other palms. American Journal of Botany 54:1143-1151.

Patel, R. N. 1965. A comparison of the secondary xylem in roots and stems. Holzforschung 19:72-79.

Pickard, W. F. & W. F. Melcher. 2005. Perspectives on the biophysics of xylem transport. Pp 3-18. In: N. M. Holbrook & Ma A. Zwieniecki (eds). Vascular transport in plants. Elsevier Academic Press, Oxford.

Pitterman, J., J. S. Sperry, U. G. Hacke, J. K. Wheeler & E. H. Sikkema. 2005. Torus-margo pits help conifers compete with angiosperms. Science 310:1924.

Rabaey, D., F. Lens, E. Smets & S. Jansen. 2006. The micromorphology of pit membranes in tracheary elements of Ericales: new records of tori or pseudotori? Annals of Botany 98:943-951.

Rudall, P. 1991. Lateral meristems and stem thickening growth in monocotyledons. The Botanical Review 57:150-163.

--1995. New records of secondary thickening in monocotyledons. IAWA Journal 16:261-268.

Sauter, J. J., VC. 1. Iten & M. H. Zimmermann. 1973. Studies on the release of sugar into the vessels of sugar maples (Acer saccharum). Canadian Journal of Botany 51 : 1-8.

Schneider, E. L. & S. Carlquist. 1998. Origins and nature of vessels in monocotyledons. 4. Araceae subfamily Philodendroideae. Bulletin of the Torrey Botanical Club 125: 253-260.

Silvera, K., L. S. Santiago, J. C. Cushman & K. Winter. 2010. The incidence of crassulacean acid metabolism in Orchidaceae derived from carbon isotope ratios: a checklist of the flora of Panama and Costa Rica. Botanical Journal of the Linnean Society 163:194-222.

Slatyer, R. O. 1976. Plant-water relationships. Academic, London. Smith, S. D., T. L. Hartsoek & P. S. Nobel. 1983. Ecophysiology of Yucca brevifolia, an arborescent monocot of the Mojave Desert. Oecologia 60: 10-17.

Solereder, H. & F. J. Meyer. 1930. Systematische Anatomic der Monokotyledonen. Heft VI. Scitamineae. Verlag Fischer, Berlin.

Soltis, D. E., P. S. Soltis, M. W. Chase, M. E. Mort, D. C. Albach, M. Zanis, V. Savolainen, W. H. Hahn, S. B. Hoot, M. F. fay, M. Axtell, S. M. Swensen, L. M. Prince, W. J. Kress, K. C. Nixon & J. S. Farris. 2000. Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133: 381-461.

Sperry, J. S. 1985. Xylem embolism in the palm Rhapis excelsa. IAWA Bulletin, new series 6: 283-292.

--1986. Relationship of xylem embolism to xylem pressure potential, stomata closure, and shoot morphology in the palm Rhapis excelsa. Plant Physiology 80:110-116.

--, U. G. Hacke, T. S. Fetid, Y. Sano & E. H. Sikkema. 2007. Hydraulic consequences of vessel evolution in angiosperms, lnternational Journal of Plant Sciences 168:1127-1139.

Stevenson, D. W. 1980. Radial growth in Beaucarnea recurvata. American Journal of Botany 67: 476-489.

Stiller, V., J. S. Sperry & R. Lafitte. 2005. Embolism conduits of rice (Oryza sativa, Poaceae) refill despite negative pressure. American Journal of Botany 92: 1970-1974.

Takahashi, A. 1988. Morphology and ontogeny of stem xylem elements of Sarcandra glabra (Thunb.) Nakai (Chloranthaceae): additional evidence for the occurrence of vessels. The Botanical Magazine (Tokyo) 101: 387-395.

Thorsch, J. 2000. Vessels in Zingiberaceae: a light, scanning, and transmission microscope study. IAWA Journal 21:61-76.

Tippo, O. 1946. The role of wood anatomy in phylogeny. American Midland Naturalist 36: 362-372.

Tnmlinson, P. B. 1961. Anatomy of the monocotyledons. II. Palmae. Clarendon, Oxford.

--1969. Anatomy of the monocotyledons. III. Commelinales Zingiberales. Clarendon, Oxford.

--1983. Anatomy of the monocotyledons. VII. Helobieae (Alismatidae). Clarendon, Oxford.

--1990. The structural biology of palms. Clarendon, Oxford.

--& M. H. Zimmermann. 1969. Vascular anatomy ofmonocotyledons with secondary growth--an introduction. Journal of the Arnold Arboretum 50: 159-179.

Turrlll, W. B. 1942. Taxonomy and phylogeny. Part II. Botanical review 8:473 532.

Tyree, M. T. & M. H. Zimmermann. 2002. Xylem structure and the ascent of sap, ed. 2. Springer Verlag, Berlin.

Van Ieperen, W. 2007. Ion-mediated changes of xylem hydraulic reistance in planta: fact or fiction? Trends in Plant Science 12: 137-142.

Vogt, K. 2001. Hydraulic vulnerability, vessel refilling, and seasonal courses of stem water potential of Sorbus aucuparia L. and Sambucus nigra L. Journal of Experimental Botany 52:1527-1536.

Wagner, P. 1977. Vessel types of the monocotyledons. A survey. Botaniska Notiser 130:119-147.

Woodhouse, R. M., 3. G. Williams & P. S. Nobel. 1980. Leaf orientation, radiation interception, and nocturnal acidity increases by the CAM plant Agave deserti. American Journal of Botany 67:1179-185.

Young, D. A. 1981. Are the angiosperms primitively vesselless? Systematic Botany 6:313-320.

Zimmermann, M. H. 1983. Xylem structure and the ascent of sap. Springer Verlag, Berlin.

Sherwin Carlquist (1, 2)

(1) Santa Barbara Botanic Garden, 1212 Mission Canyon Road, Santa Barbara, CA 93105, USA

(2) Author for Correspondence; e-mail: s.carlquist@vefizon.net

Published online: 5 April 2012

DOI 10.1007/s12229-012-9096-1
COPYRIGHT 2012 New York Botanical Garden
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Carlquist, Sherwin
Publication:The Botanical Review
Article Type:Report
Geographic Code:1USA
Date:Jun 1, 2012
Words:28490
Previous Article:Biased sex ratios in plants: theory and trends.
Next Article:Linear Trends in Botanical Systematics and the Major Trends of Xylem Evolution.
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |