Columbia University Press

Publishers Since 1893

New York Chichester, West Sussex

Copyright © 2009 Columbia University Press

All rights reserved

Library of Congress Cataloging-in-Publication Data

Stuessy, Tod F.
   Plant taxonomy: the systematic evaluation of comparative data / Tod Stuessy—2nd ed.
   p. cm.
Includes bibliographical references and index.
   ISBN 978-0-231-14712-5 (cloth : alk. paper)
   ISBN 978-0-231-51864-2 (ebook : alk. paper)
1. Plants—Classification   I. Title

   QK95.S78 2008
   580.1´2—dc22
2008028614

∞
Columbia University Press books are printed on permanent and durable acid-free paper.
This book is printed on paper with recycled content.
Printed in the United States of America

c 10 9 8 7 6 5 4 3 2 1

References to Internet Web sites (URLs) were accurate at the time of writing. Neither the author nor Columbia University Press is
responsible for URLs that may have expired or changed since the manuscript was prepared.

Stuessy_FM_3rd.indd 4 10/24/08 12:57:55 PM

 16
CHAPTER

Anatomy
Anatomy, or the internal form and structure of plant organs, is
another classical source of data used in plant taxonomy. Anatomical
data are often extremely useful in solving problems of relationships
because they can often suggest homologies of morphological
character states, and they can also help in the interpretation of
evolutionary directionality (= polarity). Comparative anatomy
is sometimes characterized as a sterile discipline, bereft of new
advances, that has outlived its usefulness. Nothing could be
further from the truth. A glance at the papers in symposia volumes
(e.g., Robson, Cutler, and Gregory 1970; Baas 1982b; White and
Dickison 1984; Rudall and Gasson 2000); the stimulating ecological
hypotheses of Carlquist (1975); and the architectural form and
function analyses of Tomlinson, Fisher, Honda, and coworkers (e.g.,
Tomlinson and Zimmermann 1978; Halle, Oldeman, and Tomlinson
1978; Honda, Tomlinson, and Fisher 1981, 1982; Fisher 1984) and
other investigators (André 2005; Turnbull 2005) reveals many new
challenges. In fact, interest has focused clearly on the function and
adaptive value of anatomical features, even in fossil plants (e.g.,
Taylor 1981; Speck et al. 2003), which will help reveal more clearly

Stuessy_ch16_3rd.indd 199 10/24/08 2:23:10 PM

 200 taxonomic data
the homologies of structure for purposes of classification and
the reconstruction of phylogeny.
History of Anatomy in Plant Taxonomy
The history of use of anatomical data in systematics is long and
follows in parallel fashion the use of more explicit morphologi‑
cal data. Good summaries are provided by Eames and MacDan‑
iels (1947), Carlquist (1969a), and Metcalfe (1979), with more
recent developments chronicled by Endress, Baas, and Gregory
(2000) and Rudall (2007). The beginnings can be traced back
to the English worker, Nehemiah Grew, especially as revealed
in his book The Anatomy of Plants (1682), and to the Italian
Marcello Malpighi (Anatome Plantarum, 1675–1679). John
Hill (1770) followed with his book, The Construction of Tim-
ber. The real development of plant anatomy came in the middle
of the nineteenth century, beginning with Hugo von Mohl’s
Grundzüge der Anatomie und Physiologie der vegetabilischen
Zelle (1851). Among the many anatomical works that followed
was L. Radlkofer’s anatomical treatment (1895) of Sapindaceae
for Engler and Prantl’s Die natürlichen Pflanzenfamilien. His
student, H. Solereder, wrote the important Systematische Anat-
omie der Dikotyledonen (1908) and set the stage for continued
presentations of comparative data ending with the compendia
of C. R. Metcalfe alone (1960, 1971) and with L. Chalk (1950,
1979, 1983). In the United States, comparative plant anatomy
developed primarily under E. C. Jeffrey at Harvard as reflected
in his book The Anatomy of Woody Plants (1917). His students,
such as I. W. Bailey, A. J. Eames, E. W. Sinnott, and R. H. Wet‑
more, all have made significant contributions. Other important
figures and their works include A. S. Foster (Practical Plant
Anatomy, 1942), K. Esau (Plant Anatomy, 1953; Anatomy of
Seed Plants, 1960; Vascular Differentiation in Plants, 1965), and
S. Carlquist (Comparative Plant Anatomy, 1961).
General Anatomical Texts
and References
Many books contain anatomical data about plants. Basic ana‑
tomical textbooks include those by Eames and MacDaniels
(1947), Esau (1953, 1960), Carlquist (1961), Molisch and Ho‑
fler (1961), Cutter (1971, 1978), Mauseth (1988), Fahn (1990),
Braune, Lehmann, and Taubert (1999), Cutler, Botha, and
Stevenson (2005), Evert (2006), and Rudall (2007). Carlquist
(1961), who dealt with comparative anatomy, offered useful
perspectives for the taxonomist. The review by Endress, Baas,
and Gregory (2000) is also extremely helpful. Books of interest
that deal with general aspects of developmental anatomy were
written by Torrey (1967), Steward (1968), and O’Brien and Mc‑
Cully (1969). Wood anatomy was covered well by Desch (1973),
Jane (1970), Panshin and de Zeeuw (1980), Zimmermann and
Stuessy_ch16_3rd.indd 200
Brown (1971), Core, Côté, and Day (1979), Zimmermann
(1983), and Carlquist (2001a). Other specific aspects of vegeta‑
tive anatomy include works on shoots and meristems (Clowes
1961; Romberger 1963; Dormer 1962; Williams 1974; André
2005), leaf development (Maksymowych 1973; Dale and Milt‑
horpe 1983), phloem (Esau 1969), cambium (Philipson, Ward,
and Butterfield 1971), general features of vascular differentia‑
tion (Esau 1965), roots (Torrey and Clarkson 1975), and plant
architecture (Turnbull 2005). Ultrastructural topics are cov‑
ered by Côté (1965), Robards (1974), and Gunning and Steer
(1975). Information on applied and economic aspects of plant
anatomy can be found in Hayward (1938) and Cutler (1978a).
Standard texts for microtechnique used in anatomical research
are Johansen (1940), Sass (1958), Jensen (1962), Berlyn and
Miksche (1976), and O’Brien and McCully (1981); for electron
microscopy techniques see Hayat (1970, 1978, 1981), Gabriel
(1982a, b), Dykstra (1992), Slayter and Slayter (1992), Cheng
et al. (1993), and Fleger, Heckman, and Klomparens (1993). The
most valuable sources for comparative anatomical data about
angiosperms are the publications of Metcalfe and Chalk (1950,
1979, 1983), which are organized by family in the Bentham
and Hooker system of classification and also have many refer‑
ences to the included genera. A useful source of data on leaf
structure of tropical trees is Roth (1984). The Bibliography of
Systematic Wood Anatomy of Dicotyledons (Gregory 1994) has
a wealth of references on this aspect. More detailed anatomi‑
cal monographs are available for some groups, especially in the
monocots (e.g., Gramineae, Metcalfe 1960; Palmae, Tomlinson
1961, 1982b; Juncales, Cutler 1969; Commelinales‑Zingibera‑
les, Tomlinson 1969; Cyperaceae, Metcalfe 1971; Dioscoreales,
Ayensu 1972; Helobiae, Tomlinson 1982b; Iridaceae, Rudall
1995). In short, a wealth of anatomical data is available for po‑
tential use as taxonomic characters.
Types of Anatomical Data
Anatomical data can be viewed as consisting of two types: en‑
domorphic (as contrasted with exomorphic, or morphological,
data), and ultrastructural. The former are observable largely
with the light microscope and the latter by use of the trans‑
mission electron microscope (TEM; often used with algae,
e.g., Watanabe and Floyd 1989; Cook et al. 1997). Another po‑
tentially very useful technique is nuclear magnetic resonance
(NMR) imaging, which allows nondestructive data gathering
and electronic “dissection” (Veres, Cofer, and Johnson 1991;
Mill et al. 2004). As one approaches comparative cellular struc‑
ture with TEM, the transition with cytology is reached. This is
especially the case with sophisticated techniques such as scan‑
ning transmission electron microscopy (STEM; Linderoth, Si‑
mon, and Russel 1997), scanning tunneling microscopy (STM;
Kolb, Ullmann, and Will 1997; Lopinski et al. 1998; Hörber
and Miles 2003; De Wilde et al. 2006), far-field optical nanos‑
10/24/08 2:23:10 PM

copy (Hell 2007), surface-enhanced raman scattering (SERS;
Nie and Emory 1997), cryo-electron tomography (Grünewald
et al. 2003; Kürner, Frangakis, and Baumeister 2005; Nicastro
et al. 2006), and use of the atomic force microscope (AFM;
Radmacher et al. 1994; Hansma et al. 2006; Sugimoto et al.
2007). All of these allow even individual molecules or atoms to
be visualized. In this book, cytological data are confined mostly
to information about chromosomes.
Numerous applications of anatomical data for solving sys‑
tematic problems exist in the literature, and it is not our task
here to chronicle all of them. The purpose of the following
discussion is to give examples of kinds of anatomical and ul‑
trastructural data that can be useful in solving different kinds
of taxonomic problems. For a good list of potential anatomi‑
cal characters and states from different plant organs, see Rad‑
ford et al. (1974).
Vegetative Anatomical Characters
In contrast to vegetative morphological features, vegetative
anatomical characters have been used with more regularity
than floral ones. This is probably due to the viewpoint that
Stuessy_ch16_3rd.indd 201
anatomy 201
if additional data are believed desirable to solve a taxonomic
problem, then looking inside the leaves, stems, and roots
could potentially yield different information than that from
reproductive organs. Data from floral and fruit anatomy usu-
ally correlate well with observed reproductive morphological
features and, hence, serve to refine the relationships already
documented instead of offering totally new insights. Vegeta‑
tive and reproductive characters, obviously, are sometimes
examined together (e.g., Hils et al. 1988; Gornall 1989).
Leaves Within leaves, data can be taken from the petiole,
blade, or cotyledons. An example of the first comes from
Schofield (1968) in Guttiferae, the second from Sajo and Ru‑
dall in Qualea and Ruizterania (Vochysiaceae; 2002; fig. 16.1),
and the third from Philipson (1970) in Rhododendron (Eri‑
caceae). See also Robinson (1969) in Bromeliaceae, Lackey
(1978) in Leguminosae, Glassman (1972) in Palmae, and
Keating (1984) in Myrtales.
Most leaf features of taxonomic significance derive from
the blade. The epidermis (and hypodermis, when present)
provide many useful characters as shown by Stace (1966),
Baranova (1972), and Cutler et al. (1980), but this topic (as
Figure 16.1 Cross sections of leaves
through midrib (A–C) and blade (D–H)
of species of Qualea and Ruizterania
(Vochysiaceae). A, G, Q. paraensis;
������
B, C, Q. parviflora; D, Q. grandiflora;
E, Q. dichotoma; F, Q. cordata; H,
R. albiflora. Scale
����������������
bars: A–C =�� 250
���� µm;
����
D–H = 150 µm. (From Sajo and Rudall
2002:357)
10/24/08 2:23:13 PM
 202 taxonomic data

micromorphology) has already been discussed in the previ‑

ous chapter. There are ultrastructural studies of epidermal cell
features, such as guard cell variations in Gramineae (Brown
and Johnson 1962), phytoglyphs in Eucalyptus species (Carr,
Milkovits, and Carr 1971), and crystals in nuclei of epidermal
cells (Speta 1977). The mesophyll offers some useful features,
including also the presence of crystals (Heintzelman and How‑
ard 1948, Icacinaceae; Mathew and Shah 1984, Verbenaceae).
The structure of the bundles can also vary (e.g., fig. 16.2; Brittan
1970; D’Arcy and Keating 1979), especially in Gramineae with
C4 photosynthesis (Kranz anatomy), in which bundle/sheath
cells have chloroplasts centrifugally localized and without
grana (Johnson and Brown 1973; Brown 1977). Patterns of ve‑
nation are also useful (e.g., Dickison 1969) as has been shown
on numerous occasions (fig. 16.3), and these studies border on
morphology (see fig. 15.2). Sclereids in leaves are taxonomi‑
cally valuable, too, as indicated by Rao, Bhattacharya, and Das
(1978; Rhizophoraceae), Tucker (1977; Magnoliaceae), and
Flores-Cruz et al. (2004; Mimosa, Leguminosae). See also Rao
and Das (1979) for the distribution of leaf sclereids within
Dahlgren’s system of classification of the angiosperms. Leaf

Figure 16.3 Leaf venation in species of Qualea and Ruizterania

(Vochysiaceae). A, ��� Q. cordata; B, H, Q. parviflora; C, F, R. albi-
flora; D, I, Q. dichotoma; E, Q. grandiflora; G, Q. paraensis. Scale
������
bars: A–C = 1 cm; D–H = 300 µm; I = 150 µm. (From Sajo and
Rudall 2002:356)

anatomical data, as with most all other types of data, have also
been treated phenetically for a more quantitative view of rela‑
tionships (e.g., Lubke and Phipps 1973).

Stems Tissues and cells of stems also provide many helpful

Figure 16.2 Leaf vascular bundles (phloem poles uppermost) in
species of Sisyrinchium (Iridaceae). A, B, S. bermudianum; C,
S. convolutum; D, S. filifolium. Scale bars = 50 µm. (From Gold‑
blatt, Rudall, and Henrich 1990:499)
lines of taxonomic evidence. Throughout the stems of both
herbaceous and woody plants are found starch grains that are
of some limited taxonomic utility (Czaja 1978). In plants that
have laticifers (or latex‑containing ducts, as in Euphorbia‑
ceae), the anatomy of these structures and their starch grains,
as well as the chemistry of the latex itself are taxonomically
important (Mahlberg 1975; Biesboer and Mahlberg 1981;
Mahlberg and Pleszczynska 1983; Mahlberg et al. 1983). At
the ultrastructural level, it has been shown that members
of the closely related Cruciferae and Capparaceae (Cappa‑
rales) have dilated cisternae of the endoplasmic reticulum
(ER) in phloem parenchyma cells filled with protein gran‑
ules, filaments, or tubules (Hoefert 1975). These features are
also known from root cap and epidermal cells and in leaves

Stuessy_ch16_3rd.indd 202 10/24/08 2:23:16 PM


(Iversen 1970a; Behnke 1977a, c), as well as in the inner in‑
teguments of ovules (Ponzi, Pizzolongo, and Caputo 1978).
Related features are ER-dependent vacuoles that also contain
proteinaceous material. More studies by Behnke (1977c) sug‑
gested that these may be structurally closely related to the
dilated cisternae. The function of the protein-containing ER
segments is not clear (Behnke 1977c), but the possibility ex‑
ists that they contain myrosinase (Iversen 1970b), an enzyme
that hydrolyzes glucosinolates (mustard oil glucosides) found
in families of Capparales.
One of the most interesting taxonomic features of stems
is the variation in sieve-tube element plastids. In a series of
papers (Behnke 1967, 1968, 1969, 1971, 1972, 1973, 1974a, b,
c, 1975a, b, c, d, 1976a, b, c, 1977a, b, 1978, 1981a, b, 1982a, b,
c, d, 1984, 1986a, b, 1988, 1993, 1995, 1997, 1999; Behnke and
Turner 1971; Behnke et al. 1974, 2000; Hunziker et al. 1974;
Stuessy_ch16_3rd.indd 203
anatomy 203
Behnke and Dahlgren 1976; Mabry and Behnke 1976; Mabry,
Behnke, and Eifert 1976; Behnke and Mabry 1977; Behnke,
Pop, and Sivarajan 1983; Eleftheriou 1984), Behnke and col‑
leagues have shown the occurrence of several basic plastid
types (fig. 16.4), with differences relating to the shape and
size of starch and/or protein inclusions. Their most signifi‑
cant result was to show that the PIII‑type plastid is restricted
to Caryophyllales (and more broadly, Caryophyllidae), which
in addition to morphological features, is also characterized by
betalains (discussed in Chapter 19). This is the best example
of the efficacy of an ultrastructural feature at the ordinal (or
subclass) level. Other plastid types also help define the Mono‑
cotyledonae (PII) and Fabiflorae (PIV). The different charac‑
ter states have also been treated in transformation series to
reflect their presumed phylogeny, with the starch type prob‑
ably being most primitive in the angiosperms (Walker 1979;
Figure 16.4 Sieve-element plastids of the angio‑
sperms. A, B, monocotyledon subtype, with
cuneate protein crystalloids; C, D, Caryophyllidae
subtype, with ring-like bundle of filaments (f) and
crystalloids (c); E, F, Fabales subtype, with polygo‑
nal crystals (c) and starch inclusions (s). (From
Behnke and Dahlgren 1976:292)
10/24/08 2:23:19 PM
 204 taxonomic data
Figure 16.5 Cross sections of palm stems. A,
B, Euterpe precatoria var. longivaginata; C, D
(under polarized light), Metroxylon sagu; E, F,
Socratea exorrhiza. Scale bars: A, C, E = 3 mm;
B, D, F = 350 µm. (From Tomlinson 2006:11)
Behnke 1988). The function of these specialized plastids is
unclear, but the P‑type plastid at least may have to do indi‑
rectly with sealing the sieve‑plate pores of injured sieve-tubes
with P‑protein. That the angiosperms alone have this type of
protein may relate to the larger size and greater penetrability
of their sieve‑area pores compared to those of gymnosperms
and lower vascular plants (Evert 1984).
Many additional, significant, taxonomic features have
been derived from stem tissue (fig. 16.5). Leaf gaps and
nodal anatomy are two of these useful features, and they can
be helpful in both herbaceous and woody plants (Dickison
1969, 1975; Howard 1970; Keating 1970). Aspects of many
different tissues in stems of rattan palms have been com‑
pared (Mathew and Bhat 1997), as have patterns of primary
vasculature in Chenopodiaceae (Al-Turki, Swarupanandan,
and Wilson 2003). The general utility of wood anatomy has
been documented extensively (e.g., Brazier 1968) and hun‑
dreds of studies exist (e.g., Carlquist 1966, 1969c, 1970, 1971,
1978, 1981b, 1982a, b, c, d, e, 1984a, b, 1985a, b, 1996, 2000,
2003; Stern 1967; Keefe and Moseley 1978; Schmid and Baas
1984; Vliet and Baas 1984; Jansen, Piesschaert, and Smets
2000; Lens, Smets, and Jansen 2004; Carlquist 2005; Lens et
al. 2005a, b). Generally, transverse (cross‑; fig. 16.6), radial,
and tangential sections are made of the secondary xylem, and
the arrangements of tissues compared and contrasted for tax‑
onomic utility. Detailed microscopic study of the individual
cell types may also prove significant (see Findlay and Levy
1970, for good SEM photomicrographs of details of wood
structure). For example, there is taxonomic significance as to
whether pits in the secondary xylem are vestured (i.e., fringed
or not; Jansen, Baas, and Smets 2000, 2001; fig. 16.7) and in
the nature of remnants of pit membranes in vessel elements
(Carlquist 1992; Carlquist and Schneider 2004). Wood fea‑
tures in fossils are of obvious value for assessing relationships
Stuessy_ch16_3rd.indd 204
especially of extinct forms (Kenrick and Crane 1997) and also
for helping interpret palaeoenvironments (Woodcock and
Ignas 1994; Poole 2000; Hughes, Swetman, and Diaz 2007).
In many cases, good correlation is seen between stem anat‑
omy and the new molecular (DNA) data (Baas, Wheeler, and
Chase 2000; Endress, Baas, and Gregory 2000; Olson, Gaskin,
and Ghahremani-nejad 2003).
Roots Roots are neglected vegetative features for taxo‑
nomic characters perhaps due partly to the difficulty of ob‑
taining materials, especially in large woody species. Also, the
data available so far reveal fewer variations in comparative
structure than with the other organs of the plant (fig. 16.8;
Raechal and Curtis 1990; Olson and Carlquist 2001); hence,
roots are regarded as less valuable taxonomically. One posi‑
tive study used root anatomy to identify eight genera of the
Caprifoliaceae found in the British Isles (Gasson 1979). No
doubt variations in root cap and meristem organization and
in mycorrhizal associations might provide more data, but so
far, these have been little explored. To emphasize this point,
Vandenkoornhuyse et al. (2002) found 49 different fungal
genotypes (from all the higher fungi) inside the roots of the
grass Arrhenatherum elatius. It is unclear what these fungi are
doing there.
Reproductive Anatomical Data
Reproductive anatomical features have also been used in
taxonomic studies to good effect (see general review by Puri
1951; see also Eyde 1975a; for a good example involving many
different characters, see Tobe, Carlquist, and Iltis 1999). The
flower contains many patterns of vascularization that are
useful in showing relationships in specific instances (e.g., in
Hamamelidaceae, Bogle 1970, and in Saururaceae, Liang and
10/24/08 2:23:22 PM
 anatomy 205

Figure 16.6 Transverse sections of woody stems in

taxa of Myrsinaceae. A, Grammadenia parasitica;
B, Ctenardisia stenobotrys; C, Embelia kilimand-
scharica; D, Geissanthus quindiensis; E, Oncoste-
mum leprosum; F, Ardisia cauliflora. (From Lens
et al. 2005b:169)

Figure 16.7 SEM photomicrographs of vestured and non-vestured

pits in secondary xylem of taxa of Malvales s.l. A, Bixa orel-
lana; B, Halimium alyssoides; C, Lechea patula; D, E, Cochlo-
spermum regium; F, G, Diegodendron humbertii; H, Marquesia
sp.; I, Monotes kerstingii. All views are of outer surfaces except
for inner surfaces in G and I. (From Jansen, Baas, and Smets
2000:174)

Stuessy_ch16_3rd.indd 205 10/24/08 2:23:31 PM

 206 taxonomic data
Figure 16.8 Woody root structure in transections
(A, D; pith below), vessels from tangential sections
(B, E), and tangential sections (C, F) in Circaea
lutetiana (A–C) and Lopezia suffrutescens (D–F) of
Onagraceae. Divisions = 10µm. (From Carlquist
1982e:763)
Tucker 1990). Features of the ovarian disc (Rao 1971), pla‑
centation (Rao 1968), gynoecium (Eyde 1967), and more spe‑
cifically the carpels (Endress 1980a, b; Endress and Lorence
1983; Endress and Igersheim 2000), including septal nectaries
(Simpson 1993; fig. 16.9), are also important. A number of
studies have covered several of these floral aspects together
(in Polygalaceae, Eriksen 1993b; Vivianiaceae, Narayana
and Rama Devi 1995; Bretschneidera, Ronse De Craene et al.
2002). The ontogeny of floral structures and vascularization
can also be helpful taxonomically (in Saururaceae, Liang and
Tucker 1995; in Koelreuteria, Sapindaceae, Ronse De Craene,
Smets, and Clinckemaillie 2000). The anatomy of the fruits
also can be examined (in achenes [= cypselae], Chute 1930,
Pandey, Chopra, and Singh 1982, Stuessy and Liu 1983, Cron,
Robbertse, and Vincent 1993; in Anacardiaceae, Wannan and
Quinn 1990, Rohwer 1995, 1997, Hofmann 1999), as well as
seeds (e.g., Tobe, Wagner, and Chin 1987; Doweld 1996; Nandi
1998b; Svoma 1998a), including protein body inclusions (Lott
1981). Comparative cotyledonary features have also been
Stuessy_ch16_3rd.indd 206
used in Acacieae (Scott and Smith 1998). A good example of
a detailed anatomical study of seeds is that by Elisens (1985)
in tribe Antirrhineae of Scrophulariaceae. Nuclear magnetic
resonance has also been used to reveal internal fruit and seed
structures (Mill et al. 2004; fig. 16.10). Inflorescences provide
a wealth of anatomical data, too, including developmental as‑
pects, such as revealed in studies of the subtribe Madiinae of
Compositae (Carlquist 1959), Uncinia of Cyperaceae (Kuk‑
konen 1967), Eriochloa and relatives in Poaceae (Thompson,
Tyrl, and Estes 1990), Allocasuarina of Casuarinaceae (Flores
and Moseley 1990), and the occurrence of extracellular cal‑
cium oxalate crystals in Araceae (Barabé et al. 2004).
Investments for Gathering
Anatomical Data
The equipment needed for anatomical work at the light mi‑
croscopic level differs from that of ultrastructural studies in‑
10/24/08 2:23:39 PM

volving the TEM. For light microscopy, one might need mi‑
crotomes, usually rotary, which can yield tissue slices several
microns thick. Finely honed steel knives or glass knives are
used for the cutting. Tissues are usually embedded in paraffin
or plastic, and the cut sections are mounted on glass slides,
dehydrated, stained, and counterstained to reveal the differ‑
ent tissues. Resin can be injected into internal spaces, and the
organic tissues dissolved to reveal casts of internal structure
(Mauseth and Fujii 1994). For TEM, an ultramicrotome is
used to cut 70–100 nm thick sections of very small pieces of
tissue (ca. 1 mm square) that are embedded in hard resins.
These sections are then cut with glass or diamond knives,
mounted on grids, and stained to increase electron density.
These techniques require voucher specimens, so that the ana‑
tomical sections can always be referable to a herbarium sheet
and, hence, to the macroscopic features of the taxon (Stern
and Chambers 1960).
The expertise, time, and cost needed to do anatomical
work require more of a commitment than that for morpho‑
logical studies. In comparison with costs for DNA sequencing
and fragment analyses, or even phytochemical data, however,
Stuessy_ch16_3rd.indd 207
anatomy 207
Figure 16.9 Cross sections of ovaries in flowers
of taxa of Haemodoraceae. A, B, Haemodorum
spicatum; C, D, Anigozanthos flavidus; E, F,
A. fuliginosus. Arrows in A, C, D, E point to
septal nectaries enlarged in B, D (toward apex
of ovary), E. Scale bars: A, C, E = 500 µm;
B = 50 µm; D, F = 200 µm. (From Simpson
1993:602)
anatomical investigations are much less expensive. Being
able to cut good sections with microtomes requires a defi‑
nite skill and patience not found in all taxonomists. Further,
the numerous other preparative steps, such as embedding,
dehydration, and staining, all require considerable skill and
experience. A good five‑week course in plant microtechnique
should be adequate to begin light anatomical studies with an
additional ten-week course for TEM. The cost for such efforts
in terms of equipment is small for light microscopic studies,
assuming a good compound microscope is already available.
Several thousand dollars will set up a rudimentary bench for
beginning anatomical work. With TEM, however, the costs
escalate tremendously with the cost of a good TEM itself be‑
ing in the range of US$100,000–200,000. Ultramicrotomes,
digital imaging equipment, and computers all require many
thousands of dollars for just the initial laboratory setting.
Fortunately, at most major institutions, a well-outfitted TEM
lab is already available, and services can be purchased for an
hourly fee or by access through training courses. The costs for
TEM work, however, are high. As mentioned in the previous
chapter, SEM is also expensive.
10/24/08 2:23:44 PM
 208 taxonomic data

Figure 16.10 Anatomical and NMR images

of seeds of species of Prumnopitys and
Afrocarpus (Podocarpaceae). A–D,�����
P. andina; E–G, P. taxifolia; H–L, P.
ferruginea; M–P, A. falcatus. A,
�����������
E, I, M
show photographs of transverse sec‑
tions (left) and NMR transverse images
(right). Most other figures are different
NMR transverse images, but H and L
are surface-rendered NMR views. avb
= ascending vascular bundle; dvb =
descending vascular bundle; e = embryo;
is = inner sarcotesta layer; ivs = inner,
vestigial, sarcotesta layer; mg = megaga‑
metophyte; os = outer sarcotesta layer;
pvb = peduncular vascular bundle;
s = sarcotesta; sc = sclerotesta. Scale bars
= 1 mm. (From Mill et al. 2004:302)

resolved by Schmid (1972a). Another is delimitation between

Efficacy of Anatomical Data in the
Taxonomic Hierarchy
Anatomical data have been used to good effect at all levels in
the taxonomic hierarchy. At the familial level and above, exam‑
ples include Baranova’s (1972) use of leaf anatomical features to
determine relationships in Magnoliaceae and related families.
Endress and colleagues have provided detailed examinations
and comparisons of different anatomical structures among
the basal angiosperms (Endress 2001b; Endress and Igersheim
1999, 2000; Alexey et al. 2005; Romanov et al. 2007). Arm‑
strong (1985) helped resolve the delimitation of Bignoniaceae
and Scrophulariaceae based on floral anatomy. The spectacular
results with sieve-tube plastids in delimitation of Caryophyl‑
lales by Behnke and coworkers (mentioned earlier in this chap‑
ter) provide the best example of efficacy of ultrastructural data
at the ordinal level. Many anatomical studies have dealt with
problems of generic delimitation, and one of the most inter‑
esting is the Eugenia-Syzygium problem in Myrtaceae that was
Menodora and Jasminum in Oleaceae (Rohwer 1997). Carlquist
(1961) and coworkers have also contributed considerably to the
solution of generic problems via floral and wood anatomy. Other
good examples are Sherwin and Wilbur (1971) in Crassulaceae
and Theobald (1967) in placement of Uldinia (Umbelliferae).
At the specific level, the contributions of anatomy tend to be
less helpful than at the generic level. As Metcalfe mentioned:
“Even when the identity of the genus has been established by
traditional methods the anatomical separation of species is of‑
ten by no means easy, especially where one is dealing with a
large genus in which the species are rather alike” (1954:434).
Anatomy can be useful at this level in some instances, however,
and even in the documentation of hybridization (Hillson 1963;
Webb and Carlquist 1964).
Of all the anatomical characters available for employment
in taxonomic studies, do some of them tend to be more effica‑
cious than others? The answer, of course, is that it will depend
on the type of question being asked and the nature of the group
under study. A useful perspective was offered by Metcalfe:

Stuessy_ch16_3rd.indd 208 10/24/08 2:23:50 PM


The investigator has to take the plants on which he is
working as he finds them and to make the best use of the
characters which they exhibit. Provided the plants under
investigation have at least one, or preferably more, of the
following attributes there are reasonable prospects that
the systematic anatomist will be able to make something
of their taxonomy. With plants that do not fulfil these
conditions he is less likely to be successful. The desirable
attributes are as follows: (a) well developed second‑
ary xylem; (b) distinctive trichomes or other dermal
appendages; (c) a characteristic distribution pattern of
sclerenchyma; (d) ergastic substances such as crystals
and siliceous bodies which are deposited in the plant
body in distinctive morphological forms. (1968:48)
Special Concerns with Anatomical Data
One of the most successful contributions of anatomy to plant
taxonomy has been in suggesting hypotheses about evolution‑
ary trends of character states. As Eames well summarized:
“These internal characters are as valuable as the external; and
some of them are perhaps even more valuable because of the
frequent persistence of the vascular supply of lost organs af‑
ter all external evidence of the organs has disappeared. These
buried vascular vestiges give information about anatomical
forms and so provide evidence of relationship to other groups”
(1953:126). Few would argue with the broadscale trends of
xylem evolution in the angiosperms, such as shown in the
evolution of vessels from tracheids (see Baas 1982b; basic
ideas originally from Frost 1930b, 1931, and Cheadle 1943a,
b), but within certain genera or families, the situation must
be interpreted with more caution. Carlquist (1969b) strongly
criticized the overly zealous attitudes of Eames and followers
and essentially concluded that floral anatomy could be ter‑
ribly misleading for the unraveling of phylogeny, despite the
fact that he himself had used these data to suggest such trends
(e.g., Carlquist 1957, 1959). This somewhat harsh and equally
exaggerated view in the opposite direction from that of Eames
has been called into question by Kaplan (1971) and Schmid
(1972b). The truth lies between these extremes (as is often the
case): floral anatomy can be extremely helpful in suggesting
phyletic sequences but only within reasonable limits. A good
example of a most helpful application is the detection of the
origin of hypogyny from epigyny in Tetraplasandra (Aralia‑
ceae; Eyde and Tseng 1969; Costello and Motley 2001, 2004),
which is the reverse of the common trend in the angiosperms
(Grant 1950; Gustafsson and Albert 1999). Here geography
provided useful ancillary data.
More recently, interest has turned to the interpretation of
floral and especially wood anatomy in a functional and ecologi‑
cal context. An important book by Carlquist (1975) hypothe‑
sized many adaptations of xylem structure with ecological con‑
Stuessy_ch16_3rd.indd 209
anatomy 209
ditions. These hypotheses have been followed by several detailed
studies (Baas 1976, 1982b; Carlquist and Bissing 1976; Carlquist
1977, 1978, 1982a, b, 1984a, b, 1985a, b; Carlquist and Debuhr
1977; Dickison, Rury, and Stebbins 1978; Rury and Dickison
1984; Davis, Sperry, and Hacke 1999; Noshiro and Baas 2000;
Cooper and Cass 2001; Gorsuch, Oberbauer, and Fisher 2001;
Feild, Brodribb, and Holbrook 2002; Landrum 2006). The re‑
sults so far confirm many of the ecological correlations with
wood structure, but other aspects seem more questionable (Baas
1986; Baas, Jansen, and Wheeler 2003). What is certain is that
many of the structural features of secondary xylem surely must
be ecologically (and/or physiologically) adaptive, and more tests
and correlations are needed to determine these tolerances and
relationships. This new information, in turn, will allow for bet‑
ter understanding of homology of anatomical character states
for taxonomic use. With the greater availability and higher re‑
liability of molecular phylogenies, it is now easier to examine
trends of evolution in wood anatomical (and other) features
and to seek ecological correlations. Endress, Baas, and Gregory
(2000) called this “ecophyletic plant anatomy.” For an excellent
example of this sort of approach in Winteraceae see Feild, Brod-
ribb, and Holbrook (2002). Functional studies of other ana‑
tomical features have also been completed, such as those on the
grass ligule (Chaffey 1994), leaf anatomy and photosynthesis in
Quercus (Fagaceae; Ashton and Berlyn 1994), stem and root an‑
atomical correlations with life-form in Moringa (Moringaceae;
Olson and Carlquist 2001), mechanical properties of stems in
two species of the cactus genus Stenocereus (Niklas et al. 2003),
and biomechanical properties of stems of lianas, horsetails, and
staminal levers in mints (Speck et al. 2003).
In conclusion, what realistically can be said about the use
of anatomical data in plant taxonomy? Bailey summarized the
proper perspective that is every bit as appropriate today as it
was over 50 years ago:
It should now be clearly recognized and freely admit‑
ted that internal or endomorphic characters are inher‑
ently no more conservative or reliable than are exo‑
morphic ones. Extensive comparative investigations
of a wide range of angiosperms demonstrate that each
morphological character tends to be relatively stable in
certain groups of plants and highly plastic and variable
in others. From a taxonomic point of view, there is
no fundamental difference between the utilization of
endomorphic, as contrasted with exomorphic, char‑
acters, except for differences in the methodologies of
obtaining data. Thus, anatomical evidence merely adds
another string to the taxonomist’s bow, and strength‑
ens the summation of evidence that may be essential
in the solution of difficult problems. In other words,
anatomical characters are inherently no more reliable
than exomorphic ones, but are susceptible to equiva‑
lent and equally valid uses (1953:121).
10/24/08 2:23:50 PM