Wood Anatomy of Akaniaceae and

Svs(>-»ifl|jc Botany (19%). 21(4): pp 607-616
Copyright 1 W by the American Society of Plant Taxonomi-t-
Wood Anatomy of Akaniaceae and Bretschneideraceae:
a Case of Near-identity and its Systematic Implications
SHERWIN CARLQUIST1
Santa Barbara Botanic Garden, 1212 Mission Canyon Road, Santa Barbara, California 93105
'Address for correspondence: 4539 Via Huerto, Santa Barbara, California 93110
Communicating Editor: David £. Ciannasi
ABSTRACT. Wood anatomy is described in detail for the single-species families Akaniaceae and
Bretschneideraceae, placed close together in most phylogenetic systems. Both families have simple perforation plates with occasional scalariform plates, alternate lateral wall pitting, helical sculpturing on vessel walls,
septate fibers with pits simple or nearly so, scanty vasicentric axial parenchyma, and wide tall multiseriate
rays composed mostly of procumbent cells. Differences between the genera are relatively few and minor
compared with resemblances; close relationship is indicated. Wood anatomy of the two families resembles
that of Sabiaceae and, to a lesser extent, that of other Sapindales. The wood is distinctive within Capparales
whether that order is defined traditionally or in the light of molecular studies. Wood data are compatible with
an interpretation that Akaniaceae and Bretschneideraceae diverged from near the base of Sapindales.
The families Akaniaceae and Bretschneideraceae,
each with a single species, have been placed next to
each other in Sapindales (including Rutales of some
authors) by Dahlgren (1980), Cronquist (1981), and
Thorne (1992); Takhtajan's (1987) treament includes
them in Sapindales with only a single family
(Melianthaceae) intervening between them. The
families differ in that Akaniaceae have hypogynous
flowers and are alleged to lack myrosin cells (which
are usually thought to contain glucosinolate compounds), whereas Bretschneideraceae have perigynous flowers and myrosin cells (Cronquist 1981).
Although Akania may lack myrosin cells, both
families contain glucosinolates (Etlinger and Kjaer
1968); some families that contain glucosinolates
apparently lack myrosin cells (Metcalfe and Chalk
1950) but all possess some portion of this chemical
syndrome. Recently, similarity in embryology between Akaniaceae and Bretschneideraceae has been
demonstrated (Tobe and Peng 1990; Tobe and
Raven 1995).
The glucosinolate-bearing families, once scattered among seven orders (Batales, Capparales,
Celastrales, Euphorbiales, Geraniales, Sapindales,
and Violales; Cronquist 1981), were included in an
expanded Capparales by Dahlgren (1980) on the
basis of certain character distributions and then by
Rodman (1991a, b), on the basis of phenetics and
cladistics. Excluded from Dahlgren's (1980) revised
Capparales were the glucosinolate families Caricaceae and Pentadiplandraceae (often placed in
Violales), as well as Bretschneideraceae (usuallyincluded in Sapindales). Drypetes, alone among
Euphorbiaceae, has been reported to possess
glucosinolates, but no author has suggested that
either Drypetes or Euphorbiaceae as a whole is close
to the other glucosinolate families. The arrangement of Dahlgren (1980) was altered on the basis of
molecular data by Gadek et al. (1992), Rodman et al.
(1993), Rodman et al. (1994) and Conti et al. (1996),
who find that a monophyletic clade is formed by
the families Akaniaceae, Bataceae, Brassicaceae,
Bretschneideraceae, Capparaceae, Caricaceae, Gyrostemonaceae, Koeberliniaceae, Limnanthaceae,
Moringaceae, Resedaceae, Tovariaceae, and Tropaeolaceae. Outgroups relatively close to these
glucosinolate families include Geraniales, Myrtales,
and Sapindales, according to the data of these
papers.
The intriguing redefinition of the glucosinolate
families (which might tentatively still be termed
Capparales) by Rodman et al. (1993), Rodman et al.
(1994), and Conti et al. (1996) invites comparison of
their cladograms with data from wood anatomy. 1
have examined wood of Bataceae (Carlquist 1978),
Brassicaceae (Carlquist 1971), Gyrostemonaceae
(Carlquist 1978), Tovariaceae (Carlquist 1985), and
Limnanthaceae and Tropaeolaceae (Carlquist and
Donald 1996). Numerous papers contain data on
wood anatomy of Capparaceae (see Gregory 1994).
Wood descriptions have been furnished far. Koeberliniaceae (Gibson 1979) and Salvadoraceae (den
Outer and van Veenendaal 1981). I plan papers on
wood anatomy of other glucosinolate families to
supplement this picture. Some information on
wood anatomy has been offered for Akaniaceae by
Dadswell and Record (1936), Heimsch (1942), Ilic
(1991), and for Bretschneideraceae by Cheng (1980),
607
608
[Volume 21
SYSTEMATIC BOTANY
Cheng et al. (1992), Heimsch (1942), Luo (1989), and
Tang (1935). The present account is offered as a way
of offering supplementary quantitative and qualitative data, using both light microscopy and scanning
electron microscopy (SEM). The material of
Bretschneidera studied by Heimsch (1942) was from
a twig, and thus may represent juvenile wood
expressions. By placing descriptions of the woods
of Bretschneidera a n d Akania together, the idea that
they are closely related, common to both traditional
phylogenetic systems as well as those based on
molecular data (e.g., Rodman et al. 1994), can be
assessed.
Data on wood anatomy are of particular interest
because there are such diverse growth forms (trees,
shrubs, stem succulents, annuals) among the
glucosinolate families. The cladogram of Rodman
et al. (1994) places Tropaeolaceae basal to Akaniaceae and Bretschneideraceae, thereby offering the
intriguing possibility, acknowledged by those authors, that there may have been marked shifts in
habit during the hypothesized divergence of the
herbaceous family Tropaeolaceae from the arboreal
sister-families Akaniaceae and Bretschneideraceae.
The nature of wood anatomy of these families
should be considered in connection with the
hypothesized shift in habit from herbaceous to
w o o d y hypothesized by Rodman et al. (1994).
Some degree of divergence of Akaniaceae and
Bretschneideraceae in wood anatomy would be
expected because of the marked geographical
separation of the t w o families. Akania bidwillii
(Hogg) Mabb., formerly known as A. hillii Hook. f.
and A. lucens (F. Muell.) Airy-Shaw, is the sole
species of Akaniaceae and is native to southeastern
Australia (Cronquist 1981). Bretschneidera sinensis
Hemsl., the only species of Bretschneideraceae,
occurs in China and Taiwan (Lu et al. 1986).
MATERIALS AND METHODS
Wood samples were obtained from xylaria (wood
sample collections). The Forestry Commission of
N e w South Wales provided the samples of Akania
bidwillii (SFCw-D10096). The Samuel J. Record
collection of the U. S. Forest Products Laboratorysupplied the wood samples of Bretschneidera sinensis (SJRw-21841, 22016). Locality data for these
samples were not available for these specimens
other than their country of origin (Australia and
China, respectively).
Sections were prepared on a sliding microtome
and stained with a safranin-fast green combination
(Carlquist 1978). A few radial sections of
Bretschneidera sinensis wood were left unstained and
were examined with SEM. Macerations were
prepared with Jeffrey's Fluid and stained with
safranin (Carlquist 1985).
Vessel diameter is measured as lumen diameter;
an attempt is m a d e to estimate the average
diameter of each vessel (e.g., the average between
widest and narrowest diameter), since that average
would be the most significant measurement with
respect to conductive capacity of a vessel. Terms are
according to the IAWA Committee on Nomenclature (1964). The ray types of Kribs (1935) have been
accepted here. The term "imperforate tracheary
elements" is used in a wide sense, including the
range from tracheids to libriform fibers, because of
the continuous nature of this range a n d because
this usage is common among recent authors on
wood anatomy. Scale bars in all Figs, represent 10
urn units.
RESULTS
Akania bidwillii (Fig. 1-8). Growth rings inconspicuous, most latewood vessels only slightly
narrower than earlywood vessels. Vessels solitary
or in small clusters or multiples (Fig. 1), mean
number of vessels per group = 1.74. Vessels
rounded in transectional outline. Vessels mostly
with simple circular to elliptical perforation plates;
a few perforation plates basically scalariform but
with numerous aberrations (interconnections between bars) present (Fig. 3). Vessel to vessel pitting
alternate, pits about 5 urn in diameter (Fig. 5).
Inconspicuous grooves interconnecting pit apertures (coalescent pit apertures) present on vessel
walls (Fig. 5, below; Fig. 6). Vessel to axial
parenchyma pitting (Fig. 8) and vessel to ray pitting
scalariform or transitional. Mean vessel diameter,
82 um. Mean vessel wall thickness, 2.5 u m . Mean
number of vessels per mm 2 , 63. Mean vessel
element length, 730 um. Imperforate tracheary
elements with very small inconspicuous borders
(Fig. 7), the borders visible in face view and in
transections. The imperforate tracheary elements
mostly septate one to three times each, and thus are
septate fibers. Mean wall thickness of septate fibers
length of septate fibers, 1488 um. Axial parenchyma
scanty paratracheal (vasicentric) with a few terminal cells at ends of growth rings. Axial parenchyma
in strands of 4-8 (mostly 5) cells. Ray cells
multiseriate (Fig. 2), heterocellular, composed
mostly of procumbent cells, with upright cells at
1996]
CARLQUIST: AKANIACEAE AND BRETSCHNEIDERACEAE
609
*
FlCS. 1-4. Wood sections of Akania bidwillii. 1. Transection; margin of growth ring across middle of photograph. 2.
Tangential section; multiseriate rays tall, wide, uniseriate rays absent. 3. Aberrant scalariform perforation plate from
radial section. 4. Transection, growth ring margin vertically down the middle of the photograph. Two rhomboidal
crystals shown in the ray cells. Figs. 1-2, scale above Fig. 1; Fig. 3, scale above Fig. 3; Fig. 4, scale above Fig. 4; divisions in
all scales = 10 urn.
610
SYSTEMATIC BOTANY
[Volume 21
>
FIGS. 5-8. Details of wood of Akania bidwillii. 5. Alternate vessel to vessel pitting from vessel in radial section; grooves
interconnecting pit apertures evident as pale diagonal streaks between pits in lower half of photograph. 6. Portion of
vessel wall from radial section to show diagonal grooves interconnecting pit apertures. 7. Pits on septate fibers from
radial section; septa out of focus above right and below left. 8. Scalariform and transitional vessel to axial parenchyma
pitting from radial section. Figs. 5-8, scale as in Fig. 3.
1996]
CARLQUIST: AKAN1ACEAE AND BRETSCHNEIDERACEAE
the tips of rays and as sheathing cells on sides of
rays (Fig. 2, arrow). Ray cells radially shorter at
margins of growth rings (Fig. 4, center). Ray cell
wall thickness about 1.8 um. Pits in ray cells nearly
all simple. Mean height of multiseriate rays, 2793
um. Mean width of multiseriate rays at widest
point, 7.1 cells (range: 3-10). Wood nonstoried.
Rhomboidal crystals are present singly in occasional ray cells (Fig. 4, arrows).
Occasional perforation plates that are nearscalariform in disposition of bars, grooves in vessel
walls, and crystals in ray cells are three features
newly reported for Akania. In other respects,
features in the above description are very similar to
those reported by Heimsch (1942). The ray type of
Akania is not really referable to any of the ray types
of Kribs (1935). Multiseriate rays are heterocellular
with elongate tips and therefore correspond to
Heterogeneous Type IIA, but uniseriate rays are
absent in Heterogeneous Type II of Kribs (1935).
This condition has been reported in Sabiaceae
(Carlquist et al. 1993).
Bretschneidera sinensis (SJRxv-22016)
(Figs.
9-18). Wood with inconspicuous growth rings
(Figs. 9, 11). Latewood vessels slightly narrower
than earlywood vessels (compare Fig. 11, below
center, with Fig. 11, above center). Vessels solitary
or in clusters, mean number of vessels per group =
1.96. Vessels rounded in transectional outline.
Perforation plates mostly simple; occasional scalariform perforation plates present, the bars sometimes
fused (Fig. 15). Lateral walls of vessels with pits
alternate or opposite (Fig. 12), about 6 u m in
vertical diameter. Helical thickenings present on
vessel walls (Figs. 13, 14). Thickenings slender,
occasionally anastomosing (Fig. 13, upper left) and
sometimes with shallow grooves between thickenings in close pairs (Fig. 14). Mean vessel diameter,
81 um. Mean n u m b e r of vessels per mm 2 , 87. Mean
vessel element length, 624 um. Imperforate tracheary elements with simple pits and septate once
or twice (Fig. 16). Mean wall thickness of septate
fibers, 6.0 um. Mean length of septate fibers, 1182
um. Axial parenchyma scant)' paratracheal (vasicentric) and terminal. Vessel to axial parenchyma
pitting scalariform. Vessel to ray pitting scalariform, transitional, or opposite (Figs. 17, 18). Rays
multiseriate and uniseriate (Figs. 9,10,11), Heterogeneous Type IIA of Kribs (1935), tip cells and some
sheathing cells of multiseriate rays upright, all cells
of uniseriate rays upright; ray cells otherwise
procumbent (Fig. 10). Ray cell wall thickness varied
(Fig. 18), mostly about 3.0 um. Borders present on
611
pits of tangential walls of ray cells (Fig. 18, arrows).
Mean height of multiseriate rays, 1654 um. Mean
width of multiseriate rays at widest point, 5.8 pm.
Mean height of uniseriate rays, 173 pm. Wood
nonstoried. N o crystals observed. Ray cells a n d
axial parenchyma cells with dark-staining contents
(Figs. 10,11).
Heimsch is correct in saying that scalariform
perforation plates are only occasional in
Bretschneidera, whereas the description of Tang
(1935) suggests that scalariform perforation plates
are characteristic in the species. Heimsch (1942)
calls the imperforate tracheary elements (which he
observed in twig material) fiber-tracheids, but I was
unable to locate any borders on pits of imperforate
tracheary elements. Otherwise, my descriptions
agree with those of Heimsch (1942). Terminal
parenchyma may not be present in the ordinary
sense; it may be vasicentric parenchyma associated
with the small latewood vessels. Borders on ray cell
pits of Bretschneidera are newly reported.
DISCUSSION
Resemblances
Between Woods of Akania and
Bretschneidera. Some features in the above descriptions are widespread in wood of dicotyledons
(e.g., scanty paratracheal axial parenchyma) and
thus are not likely to be synapomorphies for the
two families. However, other features are not
common (perforation plates predominantly simple
but scalariform plates occasional in mature secondary xylem) and are likely synapomorphies for at
least the two families, and possibly some closely
related families as well. The list of features in which
Akania and Bretschneidera w o o d s agree is surprisingly extensive. This list includes inconspicuous
growth rings, occasional perforation plates with
bars numerous but often in aberrant arrangements,
vessel to vessel lateral wall pitting predominantly
alternate, vessel to axial parenchyma pitting scalariform, vessel to ray pitting scalariform to opposite,
vessels relatively long, imperforate tracheary elements all septate fibers with borders very vestigial
or absent, axial parenchyma scanty paratracheal
(possibly with a few terminal cells in growth rings)
in strands of 4-5 cells, multiseriate rays tall a n d
composed predominantly of procumbent cells,
wood nonstoried and with sparse dark deposits in
cells.
Akania and Bretschneidera differ in minor wavs.
Vessel to vessel pitting is alternate in Akania, but is
sometimes alternate, sometimes opposite in
612
SYSTEMATIC BOTANY
[Volume 21
' 1 2 . Wood sections of Bretsehneidera sinensis (SJRw-22016). 9. Transection; growth ring terminus about i the
distance from the top of the photograph, where rays widen. 10. Tangential section; in addition to wide, tall multiseriate
rays, several uniseriate rays are visible. 11. Transection to show nature of latewood: a layer of terminal axial parenchyma
is present at the end of the growth ring, middle, just below the wide vessels. 12. Vessel-to-vessel pitting from radial
section; opposite pitting is shown. Figs. 9,10, scale above Fig. 1; Fig. 11, scale above Fig. 4; Fig. 12, scale above Fig. 3.
1996]
CARLQUIST: AKANIACEAE AND BRETSCHNEIDERACEAE
613
FIGS. 13-16. Wood details of Bretsckneidera sinensis. 13-14. Helical thickenings from vessels of radial sections
(SJRw-22016). 15. Portion of a scalariform perforation plate from radial section (SJRw-21841). 16. Septate fibers from
radial section (SJRw-22016). Figs. 13-15, bars = 10 pm; Fig. 16, scale above Fig. 3).
614
SYSTEMATIC BOTANY
[Volume 21
FIGS. 17-18. Bretschneidera sinensis (SJRw-22016), pitting on ray cells from radial sections. 17. Scalariform to
transitional pitting (horizontal axis of ray horizontal in photograph). 18. Scalariform and opposite pitting of ray cells,
and, in sectional view at left, bordered pitting on tangential walls of ray cells (horizontal axis of ray oriented vertically).
Figs. 17-18, scale above Fig. 3.
Bretschneidera. Bretschneidera has helical thickenings
on vessel walls, whereas in Akania vessels there are
grooves interconnecting pit apertures; these two
features might be homologous; however, they can
co-occur in some families of dicotyledons, such as
Asteraceae (Carlquist 1966). The pits of Akania
septate fibers are vestigially bordered, whereas I
w a s unable to find any borders on pits of septate
fibers in Bretschneidera. Bretschneidera has uniseriate
rays, Akania lacks them. Crystals were observed in
rays of Akania but not in those of Bretschneidera.
When one compares these differences in character
states (some of which are merely differences of
degree) to resemblances between the two genera,
one finds the similarities much more numerous,
a n d m a n y resemblances are features not widespread in dicotyledons, so that the resemblances are
likely not all homoplasies. The differences between
the two genera are of a kind and degree one can
easily find within a family or even within a genus
such as Meliosma (Carlquist et al. 1993). Akania and
Bretschneidera are similar even in quantitative
features.
Resemblances of Akania and Bretschneidera.
Conceding that woods of Akania and Bretschneidera,
respectively, are close to each other, does the wood
of this family pair resemble that of the other
glucosinolate families? Because relevant monographs of some of these families are not at hand,
conclusions based on an adequate data base must
await further studies. Even so, analysis will not be
easy, because differences attributable to habit (e. g.,
stem succulence of Caricaceae; herbaceousness of
Tropaeolaceae) mostly d o not indicate phyletic
relationship.
Do the wood patterns of Akania and Bretschneidera
support the idea, widely held by several leading
phylogenists, that these two genera are sapindalean
or at least not far from Sapindales? If one excludes
Sabiaceae from Sapindales, a number of wood
features common to Akania and Bretschneidera
contrast with those of most but not all Sapindales.
1996]
CARLQUIST: AKANIACEAE AND BRETSCHNEIDERACEAE
The wide, tall rays of Akania a n d Bretschneidera
differ from the rays of most sapindalean families,
which have relatively narrow multiseriate rays. A
scattering of Aceraceae and Anacardiaceae have
rays that are of medium width, b u t these do not
approach the rays of Akania a n d Bretschneidera in
width (Heimsch 1942; Metcalfe a n d Chalk 1950).
Interestingly, occasional scalariform perforation
plates, reported here for both Akania a n d
Bretschneidera, have been reported for 10 genera of
Rutaceae, t w o genera of Anacardiaceae, a n d for the
latewood of Ailanthus of the Simaroubaceae (Metcalfe and Chalk 1950). Fabaceae are now regarded
as a family of Sapindales (e.g., Thorne 1992). Such a
large family might be expected to have such a wide
range of character state expressions that matches of
character states with those of Akania a n d
Bretschneidera could be found, b u t these are more
likely homoplasies than synapomorphies with
Akania a n d Bretschneidera (e.g., helical thickenings
in vessels). Such features of those t w o genera as
septate fibers, imperforate tracheary elements with
vestigial or simple pits and scanty vasicentric axial
parenchyma can b e found widely in Sapindales,
but without phyletic analysis, one must assume the
possibility that these features are synapomorphies.
However, wood anatomy of Sabiaceae (Carlquist
et al. 1993) offers more numerous similarities than
d o other families placed in Sapindales. Sabiaceae
are often regarded as near Sapindales (e. g., Thorne
1992), so they might b e expected to retain some
features seen in Akania and Bretschneidera if this
family pair and Sapindales had a common ancestry.
One can neglect Sabia in this comparison because
Sabia is a liana a n d thus wood of the arboreal
Meliostna is more pertinent for comparison. Notable
in Meliosma is occurrence of scalariform perforation
plates, characteristic of some species, whereas
others have only simple plates and some have both
simple and scalariform perforation plates (Carlquist et al. 1993). The multiseriate rays of Meliosma
are Heterogeneous Type 11A and often wide, as in
Akania a n d Bretschneidera, and in some species of
Meliosma, as in Akania, uniseriate rays are lacking.
Hexagonal crystals, formed singly p e r cell, are
found in Meliosma, as they are in Akania. Features
more commonly found in Sapindales, such as
septate fibers (with either vestigially bordered of
simple pits), scanty vasicentric axial parenchyma,
and presence of dark-staining c o m p o u n d s in rays
and axial parenchyma, are also resemblances
between Meliosma a n d Akania a n d Bretschneidera
(Carlquist et al. 1993). If, as the cladograms of
615
Rodman et al. (1993) and Conti et al. (1996) indicate,
Sapindales m a y lie close to the glucosinolate
families (Capparales in a new and expanded sense),
Sabiaceae m a y deserve attention with respect to
features in addition to those from w o o d anatomv-
LlTERATURE ClTED
CARLQUIST, S. 1966. Wood anatomy of Compositae: a
summary, with comments on factors controlling
wood evolution. Aliso 6: 25-44.
. 1971. Wood anatomy of Macaronesian and other
Brassicaceae. Aliso 7: 365-384.
. 1978. Wood anatomy and relationships of Bataceae,
Gyrostemonaceae, and Stylobasiaceae. Allertonia 5:
297-330.
—. 1985. Wood anatomy and familial placement of
Tovaria. Aliso 11: 69-76.
— and C. J. DONALD. 1996. Wood anatomy of
Tropaeolaceae in relation to habit and phytogeny.
Sida 17: 333-342.
— , P. L. MORRELL, and S. R. MANCHESTER. 1993. Wood
anatomy of Sabiaceae (s. I.); ecological and systematic
implications. Aliso 13: 521-549.
CHENG, J.-Q. 1980. Chinese tropical and subtropical timbers,
their distinction, properties, and application. Beijing:
Forestry Publishing House.
CHENG, J.-Q., J.-J. YANG, and P. Liu. 1992. Anatomy and
properties of Chinese luoods. Beijing: Forestry Publishing House.
CONTI, E., A. LlTT, and K. J. SYTSMA. 1996. Circumscription
of Myrtales and their relationships to other rosids:
evidence from rbcL sequence data. American Journal
of Botany 83: 221-233.
CRONQUIST, A. 1981. An integrated system of classification of
flowering plants. New York: Columbia Univ. Press.
DADSWELL, H. E. and S. J. RECORD. 1936. Identification of
woods with conspicuous rays. Tropical Woods 48:
1-30.
DAHLGREN, R. M. T. 1980. A revised system of classification
of the angiosperms. Botanical Journal of the Linnean
Society 80: 90-124.
DEN OUTER, R. W. and W. L. H. VAN VEENENDAAL. 1981.
Wood and bar anatomy of Azima tetracantha Lam.
(Salvadoraceae) with description of its included
phloem. Acta Botanica Neerlandica 30:199-207.
ETLINGER, M. G. and A. KJAER. 1968. Sulfur compounds in
plants. Recent Advances in Phytochemistry 1: 59-144.
GADEK, P. R., C. J. QUINN, J. E. RODMAN, K. G. KAROL, E.
CONTI, R. A. PRICE, and
E. S. FERNANDO. 1992.
Affinities of the Australian endemic Akaniaceae: new
evidence from rbcL sequences. Australian Systematic
Botany 5: 717-724.
GIBSON, A. C. 1979. Anatomy of Koeberlinia and Canotia
revisited. Madrono 26:1-52.
GREGORY, M. 1994. Bibliography of systematic wood
anatomy of dicotyledons. 1AWA Journal, Supplement
1:1-265.
616
[Volume 21
SYSTEMATIC BOTANY
HEIMSCH, C. 1942. Comparative anatomy of the secondary
xylem in the Gruinales and Terebinthales of Wettsteins with reference to taxonomic grouping. Lilloa 8:
83-198.
IAWA COMMITTEE ON NOMENCLATURE. 1964. Multilingual
glossary of terms used in describing wood anatomy.
Winterthur: Verlagsanstalt Buchdruckerei Konkordia.
ILIC, J. 1991. CSIRO atlas of hardwoods. Berlin: SpringerVerlag.
KRIBS, D. A. 1935. Salient lines of structural specialization
in the wood rays of dicotyledons. Bulletin of the
Torrey Botanical Club 64:177-186.
Lu, S. Y., K.-S. Hu, and F.-H. FAN. 1986. Bretschneideraceae, a new family record for the flora of Taiwan.
Quarterly Journal of Chinese Forestry 19: 114-119 (in
Chinese with English summary).
Luo, L. C. 1989. Economic timbers in Yunnan. Kunming:
Yunnan People's Publishing House.
METCALFE, C. R. and L. CHALK. 1950. Anatomy of the
dicotyledons. Oxford: Clarendon Press.
RODMAN, J. E. 1991a. A taxonomic analysis of glucosinolateproducing plants. Part 1. Phenehcs. Systematic
Botany 16: 598-618.
. 1991b. A taxonomic analysis of glucosinolate-
producing plants. Part 2. Cladisrics. Systematic
Botany 16:619-629.
, R. A. PRICE, K. G. KAROL, E. CONTI, K. J. SYTSMA, and
J. D. PALMER. 1993. Nucletoide sequences of the rbcL
gene indicate the monophyly of mustard oil plants.
Annals of the Missouri Botanical Garden 80: 656-699.
, K. G. KAROL, R. A. PRICE, E. CONTI, and K. J. SYTSMA.
1994. Nucleotide sequences of rbcL confirm the
capparalean affinity of the Australian endemic Gyrostemonaceae. Australian Systematic Botany 7: 57 69.
TAKHTAJAN, A. 1987. Systema magnoliophytorum. Leningrad:
Oficina Editoria "NAUKA."
TANG, Y. 1935. Notes on the systematic position of
Bretschneidera sinensis as shown by its timber anatomy.
Bulletin of the Fan Memorial Institute of Biology 6:
153-159.
THORNE, R. F. 1992. Classification and geography of the
flowering plants. Botanical Review 58:225-348.
T/OBE, H. and C.-L. PENG. 1990. The embryology and
systematic
relationships
of Bretschneidera
(Bretschneideraceae). Botanical Journal of the Linnean Society 103:139-152.
TOBE, H. and
P. H. RAVEN.
1995. Embryology
and
relationships of Akania (Akaniaceae). Botanical Journal of the Linnean Society 118: 261-274.