Click below for larger images and slideshow more...

The genus Cercidium (Fabaceae) offers a remarkable example of systematic wood anatomy. The feature shown in these scanning electron microscope photographs is one not ordinarily treated in systematic wood anatomy—the inner (lumen) surface of vessels. This is C. microphyllum, in which the vessel walls are smooth. Within the pit apertures, one can see vestures; these are present in pits in vessels in all species of Cercidium (although they do not show in all of the photographs here). Vestured pits occur in a number of Fabaceae. The vessel wall surfaces are described in detail in a 1989 paper [ PDF ].

There are notable differences in the vessel wall surfaces among the species of Cercidium. These differences could be used as criteria to separate the species of Cercidium taxonomically. They might even be used to separate subspecies. This photo shows the vessel walls of C. floridum subsp. peninsulare, Carter 4139. The walls are smooth, although vestures show clearly within the pit apertures.

The vessel wall surfaces of C. floridum, Wolf 2981, show grooves interconnecting the pit apertures. Some pits are not associated with grooves, but two or three pit apertures are included in some of the grooves.

The vessel wall surfaces of C. floridum, Bissing 180. The vessel walls in this collection have grooves interconnecting two or more apertures. In addition, however, the wall surfaces bear structures which may be called verrucae, or warts, or whatever: they are distinct from the vestures within the pit apertures (the vestures not visible in this photograph). There is some variation within vessels and within a particular plant with respect to the wall features of Cercidium, but this is a feature that could be used systematically. Care should be taken when using a feature that varies at a specific or subspecific level to look for variations within a population, however.

The vessel walls of C. macrum bear tiny warts, which are finer in texture than the vestures of the pits.

The vessel walls of C. andicola have grooves interconnecting pit apertures. In addition, there are coarse warts on the vessel surfaces. These warts are not randomly arranged, but are denser closer to the pit apertures. In addition, some of them merge into horizontal ridges.

The vessel walls of C. praecox bear warts that are coalescent and tend to form curious ridges that fade into the vessel wall surface. These warts are distinct from the vestures, which can be seen within the pit apertures. Notice that the pit apertures is each surrounded by a ridge of coalescent pit apertures

Vessel walls of C. australis bear smooth circular collars around the pit apertures. This appearance has been called “crateriform pits.” In addition, the warts on vessel wall surfaces of C. australis coalesce into short ridges that seem to be arranged in polygons.

An apparent hybrid, C. X sonorae (Carter 5596) has minimal collarlike rings on the pit apertures. In addition, the walls bear faint ridges that are not warty.



The recognition that there were different patterns in wood in different plants, as seen with the microscope, originated with Grew, surely [ PDF ].  But the idea that differences in wood (or leaf) anatomy could be correlated with the taxonomic system had to wait for a classification system that attempted to show natural relationships, such as the Englerian system. Therefore, progress in showing similarities among genera and families with respect to the taxonomic system mostly began with Solereder’s 1885 thesis,”Über die systematische Wert der Holzstruktur.”  Solereder’s knowledge of wood anatomy, combined with anatomy of other vegetative parts of dicotyledons, became best known in the 1906 volumes under the title of “Systematic Anatomy of the Dicotyledons.  Literature in this field up to 1950 was the basis for “Anatomy of the Dicotyledons” by C. R. Metcalfe and L. Chalk.  An attempt by Metcalfe at a second edition in series of smaller volumes showed the problems: literature in this field had expanded enormously, making comprehensive summaries of the anatomy of particular orders progressively more difficult as the years advanced. 
    However, a more serious problem for systematic wood anatomy became apparent in the last decade of the twentieth century.  The title of Solereder’s thesis implies that wood has a “systematic value”—in other words, is a source of evidence for the construction of the natural system.  The 20th century ideal was that a natural system could emerge if all kinds of evidence—pollen, embryology, gross morphology, etc—could be assembled.  Fallibilities of the various proposed phylogenies were evident to the knowledgeable.  The affinities of some families and orders seemed insoluble at worst, and with only modest degrees of probability in a number of cases.  With the advent of DNA-based phylogenies using cladistic methods, the situation changed markedly.  DNA-based trees, with their high degrees of statistical probabilities, became the template that revealed how woods evolved.  An interesting example of this can be found in the case of two monogeneric families of Caryophyllales from Madagascar, Asteropeiaceae and Physenaceae [ PDF ].  Study of anatomy of these families by earlier workers included comparisons with a number of major families and orders, but did not include Caryophyllales—Asteropeia and Physena had never been thought to belong to that order until DNA evidence demonstrated that likelihood.  The wood features of Asteropeia and Physena can now be comprehended as character state changes within Caryophyllales and thereby related with a high degree of accuracy to woods of other Caryophyllales.  Systematic wood anatomy of the 20th century offered relevant comparisons only to the extent that natural systems then in existence represented good guesses rather than bad ones. 
   There is another serious problem in systematic wood anatomy that should have been evident in 1966, with the establishment of ecological wood anatomy [ PDF ], in a paper that took habit into account as well as ecology.  Ecology and habit are of overriding importance in patterns of wood evolution.  This does not mean that, say, shrubs in a particular habitat will look alike regardless of their taxonomic affinities.  Each systematic group represented in a particular habitat shows adaptation to that place independently of and differently from the other plants there.    Vessel element features are of the greatest importance, although other wood features are involved [ PDF ].  Modifying features such as succulence, drought deciduousness, etc., must be taken into account.  One can say that some systematic studies of wood anatomy offer pictures primarily of an ecological nature, as in Dubautia [ PDF ] or Crossosomatales [ PDF ], while others the role of habit is more evident [ PDF ].  However, the roles of habit and of ecology are obviously clearly interrelated, and most systematic studies of plants with a range of habits show the involvement of both factors: Eriodictyon of the Hydrophyllaceae [ PDF ]; Polemoniaceae [ PDF ]; and Calyceraceae [ PDF ].  Many papers besides my own exemplify such trends, but I’ll let the authors of those speak for themselves.  The roles of habit and ecology are meaningfully included when those who study wood anatomy either have collected the material themselves or collaborate with those who have.  Can meaningful wood anatomy be pursued without reference to these factors?  I don’t think so, but others will disagree. 
   There are, however, some studies I have done that show that information from wood anatomy does have some application to systematics.  The species of Misodendron, sole genus of Misodendraceae, mistletoes parasitic on Nothofagus are distinct enough from each other so that a key to the species on the basis of wood can be constructed.  Such a key will never be used, of course—it’s really a kind of stunt to show that systematic differences in wood anatomy at the species level exist in some genera [ PDF ].  The interesting question in such a genus as Misodendron is one I couldn’t answer: why are there these various distinctions among the species, and what do they mean in terms of evolution in the genus?
   Sometimes wood anatomy was used to point to relationships of a genus.  Certainly wood anatomy of Ticodendron showed that it was a genus related to coryloid Betulaceae [ PDF ].  The wood of Ticodendron has more numerous primitive character states than other Betulaceae, so that it represents a basal element in the clade to which it belongs and is best treated as a distinct family.  Another genus in a monogeneric family, Corynocarpus, had been placed in various parts of phylogenetic systems—thirteen different genera were thought by one person or another to be related to Corynocarpus, but when the DNA evidence was in, none of those genera or the families to which they belonged was among the candidates proposed as neighbors.  Instead, DNA evidence suggested placement of Corynocarpaceae in a newly constituted order, Cucurbitales, a placement that would not have been suggested by gross morphology.  Was wood anatomy of Corynocapaceae consistent with cucurbitalean placement?  I thought so [ PDF ].
   Wood anatomy provided evidence for uniting Stilbaceae and Retziaceae into one family [ PDF ].  Dahlgren had been the only worker to suggest this, on the basis of iridoid presence.  The appearance of flowers of Retzia has prevented recognition of such a possible relationship, but once one sees Retzia as a bird-pollinated version of Stilbaceae, the relationship is clear. The fact that I was able to collect material of Retzia and the stilboid genera in South Africa permitted the study of their wood anatomy: wood of these genera would never be found in any xylarium, they are not woody enough to qualify—although they definitely do have wood. The families Akaniaceae and Bretschneideraceae had been placed in various groups, mostly not close to each other, in pre-DNA phylogenies.  DNA evidence suggested that not only were these two families both in Sapindales, they were probably close enough so that they could be united (as Akaniaceae).  Wood anatomy tended to show this also [ PDF ], although the question of two families versus one is perhaps a matter of choice.   Similarly, I looked at wood anatomy to see whether Aextoxicaceae, with its notably primitive wood, could be placed in the same order as Berberidopsidaceae,  Berberidopsidales, as DNA evidence suggested.  I could find no real evidence against such a treatment in terms of wood anatomy [ PDF ].  When only two families are in a clade, should they be united?  Once DNA evidence placed Krameriaceae clearly as a sister group of Zygophyllaceae, with no other families on that clade, that question was one that DNA could not answer.  Evidence from various areas of study are needed in such a case to show whether Krameria, sole genus of Krameriaceae, is farther from Zygophyllaceae than any two genera of Zygophyllaceae are from each other.  When I decided to study wood of Krameriaceae, therefore, I soon realized that I had to study wood of as many Zygophyllaceae as possible.  The result showed a number of features by which Krameriaceae consistently differ from Zygophyllaceae, supporting the idea that two families, rather than one, should be recognized [ PDF ].
   Sometimes, wood anatomy is one feature that speaks (together with other features) for separation of a genus into a new family.  Setchellanthus was such a genus [ PDF ].  Had Setchellanthus been more readily available, its differences from Capparaceae might have been noticed earlier, but it remained for Hugh Iltis to obtain material of this Mexican genus and foster studies on it.  Peter Raven provided material to me for studies of Takhtajania (Winteraceae) and Ticodendron, and I was happy in the case of these two genera, as well as in the case of Setchellanthus, to be one of several authors invited to survey these distinctive genera.
   DNA-based phylogenetic trees must now be the basis for selection of materials for studies in wood anatomy, and for interpretation of how wood evolves.  Thus, the ground has shifted under wood anatomy greatly since the days of Solereder.  The reason for systematic wood anatomy is no longer to offer evidence toward a phylogenetic system, but to show how wood evolves.  The papers on systematic wood anatomy prior to development of DNA-based trees certainly offer valid wood data, but we must now use it differently.  Equally necessary in the understanding of wood anatomy is the inclusion of facts from ecology, habit, age of plant, and location within the plant of the wood sampled.  Systematic wood anatomy is now evolutionary wood anatomy and ecological wood anatomy.  In my own career in this field, I had to respond to these changes.  Science changes, and while one can pursue some lines of study as though nothing has changed, that merely makes for anachronism and offers little to the advance of science.  When one compares woods, one advances science only by asking and answering questions….