Click below for larger images and slideshow more...

An SEM photograph of a perforation plate of Aextoxicon punctatum. Although some perforation plates of this species show considerable retention of pit membrane remnants, sectioning may have stripped away virtually all traces of primary walls from this particular perforation plate. Next to the vessel are some tracheids with bordered pits. Although the bordered pits may appear sparse, the area of the pit membranes on the tracheid pits is really relatively large. Collectively, the conductive area of these pits is appreciable, especially when one considers that the tracheids have narrow water columns operating through these pits. Vessels do not group in Aextoxicon, an indication that the imperforate tracheary elements are tracheids, not fiber-tracheids.

Part of the end wall of a Tetracentron tracheid, as seen in an SEM micrograph. Numerous tiny pores occur in the pit membranes. This is hardly different from the first of the perforation plates shown for Illicium cubense!

Pits in the end wall of Tetracentron tracheids, at higher magnification, to show the circular nature of the small pores. Pore size can vary in end walls of Tetracentron tracheids.

A radial section of the wood of Illicium floridanum, showing two complete perforation plates. Although not readily visible at this magnification, there are vertical strandlike remnants of the pit membranes in the perforations.

A portion of a perforation plate, seen from the outside of a vessel element, of Illicium tashiroi. The pit membrane remnants take the form of strands that run in an axial direction. Although these strands are readily visible with an SEM, one can see these strands in sections of Illicium woods with a light microscope, provided that the staining is sufficiently dark.

SEM view of a portion of a perforation from a vessel of Ascarina philippinensis. The pores in the pit membranes of the perforation plate are prominent, but occupy perhaps less that half of the areas of the pit membranes.

Portion of a perforation plate of Euptelea polyandra. The fine axially-oriented strands (with some flakes) in the perforations would not be visible with light microscopy, but are clearly resolved with SEM.

A perforation plate of Sarcandra glabra shows fine axially-oriented strands in the perforations.

Pit membrane remnants occur as a series of parallel threads in this perforation plate of Ascarina solmsiana.

In Saururus chinensis, pit membrane remnants occur only at the narrow lateral ends of perforations of scalariform perforation plates.

A perforation plate of Ascarina maheshwarii. Pit membrane remnants are present in the lateral ends of the perforations.

There are relatively few bars per perforation plate in Berzelia (Bruniaceae). The pit membranes in the perforations form a rather open pattern of microfibrillar webbing.

Pit membrane remnants form a flakelike pattern in the perforations of Hedyosmum brenesii (Chloranthaceae) vessels. This pattern is seen in other species of Hedyosmum as well.

An SEM view of the transition between a perforation plate (top) and a lateral wall area (bottom) of Hedyosmum luytenii. The presence of various degrees of pit membrane remnants in pits (alternative, perforations) of the transition shows that in primitive vessels, perforation plates often cannot be well defined.

In this perforation plate portion of Illicium cubense, pit membranes are very nearly intact, with small pores. They are amazingly like the pit membranes in tracheid end walls of Tetracentron shown above.

A perforation plate of Illicium cubense in which the pit membrane remnants are flakelike. Notice the transition to lateral wall pitting; the pit at bottom has no hydrolysis of the pit membrane and may be considered to belong to the lateral wall rather than to the perforation plate.

A portion of a perforation plate of Illicium cubense. Some of the tearing of the pit membranes may be as a result of handling or drying procedures. However, notice that the remnant strands are all oriented axially and have tiny holes in them: these appearances are probably mainly natural. Study of pit membrane remnants in perforation plates inevitably requires analysis with respect to artifact formation.

This perforation plate of I. cubense has only a few shreds and lumps adjacent to the perforations that give evidence of what happened to the pit membranes in the perforation plate. This plate represents the more typical condition in the wood studied for this species. The action of the conductive stream in producing this result should be suspected for this appearance.



   The idea that vessels with scalariform perforation plates are primitive is an old one—traceable to Bailey and Tupper (1918) certainly.  The most important contribution of I. W. Bailey in wood anatomy was probably to search for the most primitive expressions in wood anatomy in dicotyledons.  These primitive expressions occurred in more than one family, so the search was not just finding “the primitive wood.”  In the process of searching for primitive character states in wood of dicotyledons, Bailey discovered dicotyledons he thought to be primitively vesselless: Amborella, Tetracentron, Trochodendron, and all Winteraceae.  He thought that Sarcandra of the Chloranthaceae lacked vessels, but it definitely has them [ PDF ].  The reason that Bailey didn’t find them is that he relied on material from herbarium sheets—had he collected more material, and collected it in the field (I collected the woodiest specimens I could find), as I did, he would have found the vessel elements.  Also, I used a scanning electron microscope, which could easily demonstrate whether pit membranes are present in end walls of what might be either tracheids or vessel elements.  To a large degree, my studies of primitive vessels are based on scanning electron microscopy.  Scanning electron microscopy (SEM) was not available to Bailey or his student Frost, so the matter of the primitive vessel in dicot woods was left in limbo by them by virtue of the limitations of the light microscope.  The same applies to Cheadle’s work on monocots.  The nature of pit membranes in pits or of pit membrane remnants in perforations cannot be determined at all accurately by means of light microscopy.  Light microscopy does not resolve fine details in pit membranes.
   Perhaps the outstanding but not widely heralded finding by I. W. Bailey about vessel origin was the similarity between primitive vessel elements, like those of Aextoxicon [ PDF ], and earlywood tracheids of  Trochodendron and Tetracentron.  A number of Winteraceae that do not have marked growth rings also have scalariform pitting on end walls of tracheids [ PDF ].  The step from the scalariform pitting on end wall tracheids of these vesselless dicotyledons to the scalariform perforation plates of dicotyledons with numerous primitive wood features seems only a small one when studied with light microscopy.  The distinction between vessel elements and tracheids in woody angiosperms also involves a differentiation in diameter between the two cells types (such a distinction is usually not available in monocotyledons, which rarely have tracheids intermixed with vessel elements, although that does occur in some monocotyledons such as Borya). 
    With SEM, the story becomes more interesting.  The pioneering work (often overlooked) in this regard was by B. A. Meylan and B. G. Butterfield, whose book “The Structure of New Zealand Woods” (1978), has wonderful SEM illustrations of perforation plates that show “microfibrillar webs” in Ascarina, Carpodetus, and Quintinia.  I have used the term pit membrane remnants.  Although some pit membranes are indeed weblike, more often they take on other appearances.
    My work on Sarcandra led me to study the other genera of Chloranthaceae: Ascarina [ PDF ], Hedyosmum [ PDF ] and Chloranthus [ PDF ].  These genera show admirably the retention of pit membranes to various degrees.  There is very little difference between pit membranes or pit membrane remnants in some Chloranthaceae and the pit membranes of end walls of Tetracentron, which are sheets perforated by small holes easily visible with SEM.  The same is true of Bubbia of the Winteraceae [ PDF ] and Amborella (wood-anatomy-of-the-new_2001) [ PDF ]. In turn, I looked with SEM at numerous genera of dicotyledons with scalariform perforation plates, and discovered that pit membrane remnants characterize a number of these genera.  Such remnants had not been previously reported in most of the genera in which I found them.  My review of this phenomenon [ PDF ] can doubtless be expanded.  But I included SEM illustrations of pit membrane remnants in papers on Degeneria [ PDF ], Ticodendron [ PDF ], Schisandraceae [ PDF ], Illicium [ PDF ], Hydrangeales [ PDF ], and Crossosomatales [ PDF ] as well as other families for which pdfs are not shown here, such as Clethraceae, Ericaceae, and Sarraceniaceae (see list of publications in Biography and Publications section of this website). 
   What the instances of pit membrane remnant occurrence in scalariform perforation plates of dicotyledons showed was that the line between tracheids and vessel elements was—not a line at all.  The main criterion of vessels elements—that they have no pit membranes in end walls—is in fact, not available for such genera as Ascarina and Illicium.  Can one make a distinction?  For example, if more than half of a perforation area is holes in the primary wall rather than area of the residual primary wall, can one say that vessels are present?  Ironically, a better criterion is present in the fact that vessels are wider in diameter than the tracheids (or other imperforate tracheary elements) that they accompany.  But that can’t be applied in ferns or monocots (in these, xylem may consist wholly of vessels or wholly of tracheids, whereas in woody dicots, primitive vessels are always accompanied by tracheids).  But back to the end wall of vessel elements—compare the first end wall of an Illicium cubense vessel shown here with the lower-power figure of the end wall of a Tetracentron tracheid shown here.  Is there really a difference?
    The matter of a definition is not really at all important—or it would not be if those who did cladistic work did not make it an issue.  Those workers wanted to score vessels absent (0) versus vessels present (1) in their data matrices, and thus intermediacy was not a welcome thought to them.  But intermediacy is the fascinating and in fact, serendipitous finding.  What it means is that for families such as Chloranthaceae, Illiciaceae, Trochodendraceae, and Winteraceae, there is little difference between vessel elements and tracheids.  Therefore, cladistic concepts about vessels originating and vessels disappearing in clades including these families are really misleading.  Slight shifts in degree of hydrolysis of pit membranes in end walls of tracheary elements (or degree of cellulose microfibril deposition in them) may happen very easily, and indeed one sees various degrees of pit membrane presence in perforations within single wood samples of Illicium or Ascarina.  Do minor shifts in such degrees constitute a “change in character state?”  Degree of presence of pit membranes or pit membrane remnants in end walls may be morphogenetically closely linked to wideness of a tracheary element, so that formation of vessels no wider than the tracheary elements they accompany may be impossible.  Occurrence of vessel elements the same diameter as tracheids is not evident in basal angiosperms studied to date (except perhaps for some monocots!), although sometimes vessel elements and tracheids may differ in diameter by a relatively small amount, as in Sarcandra [ PDF ] and there can be overlap in diameter between the narrowest vessels and the widest tracheids in woods that have very narrow vessels, such as Grubbia rosmarinifolia.
   The presence of degrees of intermediacy between tracheids and vessel elements, if unsettling to those who deal with cladistics, is the positive evolutionary finding here.  If vessel elements originated from tracheids, the occurrence of intermediate stages should be expected.  The fact that the intermediate stages are preserved in living angiosperms is very interesting.  The fact that primitive vessel elements occur not in one or two, but in a number of families of angiosperms is curious.  The efficiency in conduction of the vessel and the safety of the tracheid probably represent little division of labor in woods such as Aextoxicon, Ascarina, or Illicium, compared to the division of labor between tracheids and vessel elements with fewer bars per perforation plate, as in, say, Empetraceae [ PDF ] or Vaccinium, not to mention the huge number of genera with simple perforation plates.  In terms of the biomass of the angiosperms as a whole, those genera with highly primitive vessel elements are small in species numbers and in their abundance within a flora.  Those genera mostly grow in places where moisture fluctuations are minor.  In other words, the “breakthrough” that vessel origin represents may not be so much in the opening stages, but in the stages where simplification of the perforation plate is already well advanced.  This idea receives support from work by Uwe Hacke.  Martin Zimmermann advanced the idea that perhaps bars in perforation plates sieve out air bubbles that freeze when water in vessels that is frozen thaws.  However, this idea has difficulties.  Although some genera with scalariform perforations plates with numerous bars grow in areas that freeze (Cornus, Hamamelis), probably most of the species with such perforation plates grow in areas that do not freeze (Canellaceae, Eupomatia, Goupiaceae, Illicium, Myristicaceae, Saurauia, Trimeniaceae and many others), and analysis of high latitude and high elevation floras shows that fewer than half of the woody species have scalariform perforation plates in vessels.  This idea would not explain perforation plates, such as those of Magnolia, that have only a few bars. 
   Pit membrane remnants: vestigial or functional?  The patterns of pit membrane remnants in perforation plates illustrated thus far suggests that they are more vestigial than functional.  One notes the various degrees of pit membrane presence in perforation plates within a single wood section of particular species, such as Illicium cubense, I. floridanum, or I. anisatum. Certainly the pores in pit membranes of “perforations”, where they occur, are smaller than the pores in the margo of a conifer tracheid, and they’re too small to pass air bubbles—but Zimmermann thought that clear perforations in a perforation plate with numerous bars were narrow enough to trap air bubbles.  One is tempted to conclude at this point that the degree of pit membrane remnant in perforations plates like those of Ascarina or Illlicium represents to a greater or lesser degree the activeness or slowness of the conductive stream.  A more active conductive stream might remove pit membranes more effectively. 
    Pit membranes of end walls of tracheids in vesselless angiosperms have small pores (Amborella, Bubbia).  This is distinctive if one compares them to the well-developed torus-margo structure of conifer pits on both end walls and lateral walls of conifer tracheids.  The margo pores are wide enough to enhance conduction; should air bubble invasion threaten the tracheid, the pit closes down by means of aspiration: displacement of the torus against the pit cavity is necessary.  In angiosperm woods, pit membranes of lateral wall pits do not have apparent pores at all at most magnifications. And the scalariform conformation of pits makes them unsuitable for the torus-margo sytem that conifers have.  In order to aspirate a pit, a circular torus swinging on threads of a circular margo can block the circular pit aperture.  A scalariform equivalent is difficult to imagine. Angiosperms tend to have stiff pit membranes, lacking tori, with small pores or no pores at all.  (Yes, tori have been found in vessel pits of a few angiosperms, but angiosperms just don’t have the conifer system).  Conifers are sufficiently successful with their tracheid structure that very little diversification has occurred in the plan (merely the invention of vessels in Gnetales).  Angiosperms apparently began with scalariform pitting on wider tracheids and vessel elements, and the conifer-type tracheid pit was not available to them.  Other mechanisms for achieving conductive safety are present in angiosperms: stiff pit membranes, small pit size, presence of tracheids as a subsidiary conductive system in a vessel-bearing wood, grouping of vessels, shortness of vessels so as to tend to confine air bubbles, etc.  Presence of pit membrane remnants is probably not a conductive safety mechanism.  The occurrence of pit membrane remnants is sufficiently scarce in angiosperms at large so that one must consider the feature vestigial with minimal adaptational significance.
    Pit membrane remnants in perforation plates: the diversity and what it means.  When one looks at pit membrane remnants, one finds that they are not all alike.  In a few, such as Ascarina philippinensis, the pit membranes can be sheets with holes of various sizes in them, and one concludes this is a little altered condition—either that or something has inhibited removal of the pit membrane.  In some, as in Illicium tashiroi or Euptelea polyandra, there are parallel threads running in an axial direction.  One imagines that such threads might represent minimal obstruction to the conductive stream, and that perhaps the conductive stream may have left those strands as the vessel began to function.  A very common pattern—if pit membrane remnants are present at all--is the presence of pit membrane remnants at the lateral ends of the perforations (Ericaceae).  This pattern probably also represents a minimal impedance configuration.  One imagines that the webs or strands at the ends of the perforations are a denser collection of microfibrils than those in the middle, which were swept away as the vessel element matured.  There may have been no such microfibrils in the centers of the perforations.  Bruniaceae and Sarraceniaceae tend to show relatively large holes in pit membranes, so that what appears is the microfibrillar network.   Flakelike pit membrane remnants (in addition to microfibrillar strands interconnecting the flakes) are shown here for Hedyosmum, and may also be seen in the Euptelea illustrated.  Two of the illustrations of Illicium cubense suggest flakelike remnants.
   Butterfield and Meylan in an important but brief paper (book chapter) in 1982 stated that hydrolysis dissolved the wall matrix materials in perforation pit membranes during vessel element maturation, leaving cellulosic strands.  The pattern of cellulosic strands remaining was the result of what the conductive stream left behind.  That certainly is an interpretation that one can accept.  There is, however, a residual question as to how extensive a pattern of microfibrils was laid down in the formation of primary cell walls of vessel elements to begin with.  Do perforation plates in some species have fewer microfibrils in the perforations prior to hydrolysis than the perforation plates in other species?  One tends to believe the latter, because so many angiosperms with scalariform perforation plates have vessel elements that, when mature, lack an evidence of pit membranes in the perforations.
   Note that the above observations on vessels of monocots are included in the “Fern and Monocot” section of this website.