Click below for larger image.



   How does a wood evolve so as to contain both tracheids and libriform fibers (or both tracheids and fiber-tracheids), rather than only tracheids, only fiber-tracheids, or only libriform fibers as the “background tissue” (imperforate tracheary element type)?  Nobody had ever considered this question, as simple and logical as it sounds.  Funny how some important questions in a field are not asked!  My point of departure was something called vascular tracheids.  They form at the ends of growth rings, where vessels become so narrow that they no longer have perforation plates in the end walls and thus are, by definition, tracheids.  If the vascular tracheids, instead of being just at the ends of the growth rings, could spread into all parts of a growth ring (in a plant with libriform fibers, say), and those tracheids were in contact with vessels, then that would be a path of origin for vasicentric tracheids.  Or, alternatively, the pathway might be a capability to form bordered pits in imperforate tracheary elements adjacent to vessels, a kind of vessel dimorphism.  This appears to happen in rosemary (Rosmarinus).  The family to which Rosmarinus belongs, Lamiaceae, has no tracheids in virtually all of the species for which wood is known—Rosmarinus is the only genus reported to have vasicentric tracheids.  The primitive condition of woods of the family can reasonably be hypothesized to be a wood with vessels and libriform fibers, but no tracheids.  The same can be said of Asteraceae, in which vasicentric tracheids are found (sparsely) in a few genera of very dry habitats, such as Artemisia, Baccharis, Chrysothamnus, and Haplopappus.  Wood of Asteraceae mostly has only libriform fibers as the imperforate tracheary element type.  Another chaparral genus of California, Dendromecon of the  Papaveraceae, has vasicentric tracheids plus libriform fibers.  Woods of Papaveraceae other than Dendromecon are characterized by libriform fibers only.
    However, I found some woods that had both tracheids and fiber-tracheids: blackberry (Rubus) and greasewood, Larrea tridentata) for example;—or even vasicentric tracheids plus libriform fibers (the chaparral species of Prunus).  So I developed the idea of tracheid dimorphism to account for production of two cell types instead of one.  Seems like a simple enough concept.  And I think it may account for most instances of vasicentric tracheids, in fact. A family such as Rosaceae is suggestive.  Woods of the family are characterized by tracheids.  Only a few, such as those chaparral Prunus species, have (vasicentric) tracheids plus libriform fibers, so the ancestral woods of Rosaceae seem to have had tracheids only. 
    Some plants may have begun with woods consisting only of fiber-tracheids as the imperforate tracheary element type (relatively few woods have fiber-tracheids) and through dimorphism in those cells, produced some vasicentric tracheids.   In the 1988 paper on tracheid dimorphism, I hypothesized a pathway different from that in Asteraceae and Lamiaceae. The shrubby species of Arctostaphylos from western North America have vasicentric tracheids plus fiber-tracheids, suggesting fiber-tracheid dimorphism.
    The 1988 study gave a phylogenetic dimension to tracheid presence in woods.  The astonishing thing is how pervasive tracheids are in the shrubby flora of the American southwest—“true” tracheids, vasicentric tracheids, or vascular tracheids characterize most chaparral shrubs (see the Carlquist and Hoekman 1985 paper [ PDF ]).   Several pathways have led to these cells, which confer conductive safety during the dry season, but the obvious value of various kinds of tracheids to these plants is clear.