The genus complex Danais-Schismatoclada-Payera (Rubiaceae). Character states, generic delimitation and taxonomic position

52 Pages·1993·28 MB·

by  R Buchner

Mus.
Bull. natl. Hist, nat., Paris, 4^ ser., 15, 1993,
Adamonia,
section B, 1-4 23-74.
n"''
:
The
genus complex
Danais-Schismatoclada-Payera
(Rubiaceae)
Character
and taxonomic
states, generic delimitation position
&
BucHNER
R. C. Puff
Summary
Investigations of various character prove Madagascar-centred
states clearly that the
:
genera Danais, Schismatoclada and Payera (incl. Coursiana) are very closely allied. The genera
had
previously been associated with different tribes, namely the Hedyoiideae {Rubiaceae subfam.
Rubioideae) and the Cinchoneae (subfam. Cinchonoideae), Individual character
states are
described and discussed, and characteristics that can be used to distinguish the genera are given.
Agreements and overlaps in certain features, that sometimes make difficult to keep genera
it
apart, are thought to be an expression of their close alliance rather than the result of convergent
An
evolution. analysis of the character states suggests that the genera in question can neither be
clearly and objectively assigned to subfam. Rubioideae (tribe Hedyotideae) nor placed in subfam.
Cinchonoideae because the genus complex contains features which, in part, are typical for one
and, in part, characteristic for the other subfamily. The genera seem to provide further evidence
for the view that the delimitation of the two subfamilies not entirely clear-cut, and that there
is
might be a systematic ''grey zone" between the two subfamilies. Although there some
is
new
justification in tentatively creating a tribe for the genera (without assigning to any of the
it
more
subfamilies), considered reasonable to refrain from doing so and rather wait
it is — until
A
more become
detailed, comparative data available. survey of the taxa appended
is it
;
includes several new combinations and the description of two new species of Payera.
Resume L'etude de plusieurs etats de caracteres montre clairement que les genres de la region
:
tnalgache Danais, Schismatoclada et Payera (incl. Coursiana) sont tres voisins. Ces genres ont ete
anterieurement ranges dans differentes tribus, a savoir les Hedyotideae {Rubiaceae subfam.
Des
Rubioideae) et les Cinchoneae (subfam. Cinchonoideae). etats de caracteres individuels sont
pouvant pour
decrits et discutes, et des caracteristiques etre utilisees distinguer les genres sont
chevauchements pour
donnees. Les concordances et les certaines particularites, qui quelquefois
comme
rendent separation des genres, sont consideres Texpression de leur proche
difficiles la
parente plutot que resultat d'une evolution convergente. L'analyse des etats de caracteres
le
suggere que genres en question ne peuvent, ni etre clairement objectivement attribues a
les et la
dans
sous-famille des Rubioideae (tribu Hedyotideae)^ ni places la sous-famille des Cinchonoi-
deae, parce que complexe generique contient des particularites qui, en partie, sont typiques
le
On
pour
pour Tune en caracteristiques Tautre sous-famille. trouve dans ces genres des
partie,
et,
preuves complementaires confirmant Fopinion que la dehmitation des deux sous-families n'est
pas tout a fait nette, et qu'il doit exister une « zone d'ombre » systematique entre elles. Bien quMI
legitime de tenter de creer une nouvelle tribu pour ces genres (sans la rattacher a aucune des
soit
sous-families), semble plus raisonnable de s'abstenir d'agir —ainsi et plutot d'attendre que des
il
Un
sommaire
donnees comparatives plus detaillees deviennent disponibles. expose des taxa
presente en annexe; inclut plusieurs nouvelles combinaisons et la description de deux
est
il
nouvelles especes de Payera.
&
Rennweg
Ralf Buchner Christian Institute of Botany, University of Vienna, 14, A- 1030
Puff,
Vienna, Austria.
1-
24
INTRODUCTION
I.
The aim of the present study was to investigate a broad spectrum of character states
initial
of Danais species in order to obtain detailed data for an evaluation of the genus' correct
Up
much
placement. to now, there has been debate about the taxonomic (tribal and
Taxonomic
subfamilial) position of Danais (see History, below, for details).
During these investigations however, soon became obvious that the generic delimita-
it,
tion of Danais not as clear-cut as thought previously. was noted that there are sometimes
is It
considerable problems in separating Danais and the Malagasy endemic Schismatoclada which,
like the former, of uncertain tribal position. Moreover, surfaced that the monotypic
is it
Malagasy
Coursiana (tribe Cinchoneae, subtribe Cinchoninae), yet another endemic, in turn
is
very close to the pair Danais-Schismatoclada, Finally, a by-chance discovery brought to light a
previously never thought-of obvious relationship between the Malagasy endemic Payera (tribe
HecJyotideae, subfam. Rubioideae) and the three before-mentioned genera-
Consequently, the aims of the study changed Next to Danais, other genera of this
all
:
obviously allied complex had to be dealt with. The problems regarding their generic
delimitation had to be tackled, and the question of their taxonomic status tribal and
(i.e.,
needed be
subfamilial position) to resolved.
AND METHODS
MATERIAL
IL
^ i -
-, I
Material of Danais, Schismatoclada, Coursiana and Payera from the following herbaria was studied
W
WU :
MO,
BM, BR, K, M, P, UPS. and (abbreviations according to " Index Herbariorum", Holmgren
A
et al., 1990). total of nearly 900 sheets was seen; it is important to note that not all material of
P
Schismatoclada in herbarium was available for study.
%
FPA
Moreover, samples preserved in or 70 ethanol were available of several taxa for detailed
morphological and anatomical
investigations.
WU,
were prepared from
Illustrations the following vouchers (deposited at unless an other
=
herbarium abbreviation given; from Madagascar unless stated otherwise; * preserved material)
is
:
—
Homolle (MO)
Danais aurantiaca — Gentry 11271 (Fig. 10, D). D. coronata (Pers.) Steud. Decary
: :
(WU)*
10767 (P) (Fig. 11, r-G). D.fragrans (Lam.) Pers. Mauritius Gueho s.n. (Fig. A-C; E-
1, 5,
: :
A A-Q;
F;6;7, A-D. F-K; Madagascar 1217 E; A-D\
8, 15, Benoist (?) (Fig. 1, 12, Jardin Botanique
:
;
&
Lowry (MO) (MO)
Tana 3858 (P) (Fig. 15, D-F), — Randrianasolo 4422 (Fig. 13, A-D), Schatz 1389
A
(Fig. 4, fr—om colour slide Schatz). D. humblotii Homolle Puff et al. 850824-1/9 (WU) * (Fig. 2, A-
:
Cows
B, 8, C-D). — D. ligustrifolia Baker 4451 (P) (Fig. 17, A-B), Decary 14191—(P) (Fig. 10, C; 11, ^-
:
(WU)*
D; G). D. microcarpa Baker Puff et 850824-1/7 (Fig. H). D. pubescens Baker
12, al. —
1,
;
:
(WU)*
Baron 1375 (P) (Fig. 12, E), Puff et al. 850824-1 j8 (Fig. G; \1, F). D. rhamnifolia Baker
\, —
:
Decary 5055 (P) (Fig. A), Perrier de la Bdthie 12593 (P) (Fig. 15, /; 17, C-D). D. volubilis Baker
1,
1
:
Bosser 6616 (P) (Fig. A-B), Jardin Botanique Tana 3861 (P) (Fig. H-I), Puff 850808-1/1
10, 15, et
al.
(WU)* (Fig. 1, D-F; 5, A-D, G-H; 8, B; 13, £; 15, G; Fig. 4, B-C from colour slides Puff mad- 1781 and
mad-
1782).
—
SF
Payera bakeriana (Homolle) Buchner Puff 22640 (P) (Fig. 14, B-Q. P. coriacea (Humbert)
«fe
:
& & &
Cows
Buchner Puff Humbert 22762 (P) (Fig. 16, Cj, Humbert 23823 (P) (Fig. A-B, D). Miller
16,
:
25
—
&
Lowry (MO)
4151 (Fig. 2, E-F). P. decaryi (Homolle) Buchner Puff Decary 10839 (P) (Fig. E-
14,
:
(WU)* D;9; D &
F), Puff 800808-1/3 —(Fig. 2, Fig. 4, from colour slide Puff mad-0491), Schatz N—icoll
&
(MO)
1236 (Fig. 10, I-K). P. glahrifolia Leroy ex Buchner Puff Cours 4930 (P) (Fig. E). P.
4,
:
&
homolleana (Cavaco) Buchner Puff Cours 2756
(P) (Fig. D).
14,
:
SchismatocJada psy—chotrioides Baker Gentry 11856 (MO) (Fig. 2, C), Schatz et al. 1376 (MO)
:
&
E-H; (MO)
(Fig. 10, 14, A). S. viburnoides Baker Croat 28889 (Fig. 16, G), Miller Lowry 4121
:
(MO)
(Fig. E-F).
16,
Methods
used agree with those described in Puff et (1993) except for the following supplements
al.
and
additions
:
SEM
For investigations, microtome and hand sections, samples from herbarium material were
first
%
%
reconstituted in a mixture (6 1) of 10 aqueous bis (2-ethylhexyl) sulfosuccinate sodium salt and 98
:
acetone for ca. 6 to 24 hours (modified from Peterson et 1978).
al.,
TEM
For studies of pollen and seeds, samples from herbarium specimens were embedded Spurr's
in
mixture (Spurr, 1969) and then ultra-thin sectioned (90 nm). Sections were contrasted with uranylacetate
and lead citrate. The polysaccharide test was carried out manually according to Thiery (1967) and Weber
(1989), respectively.
LM
For
investigations of pollen, grains were transferred to a mixture of methyl green and glycerine
Wodehouse made
gelatine according to (1935) measurements were on pollen grains so prepared. For the
;
FPA
determination of pollen nuclear numbers, pollen grains preserved in were stained with acetocarmine.
.
Notes on names used in the present article In order to avoid confusion, only accepted names
: <
"Taxonomic
used in the chapters following the History" section {Danais see revision of the genus, Puff
:
&
Buchner, press; Schismatoclada and Payera see Appendix). Only absolutely necessary, the
in if
:
name
original generic association of a specific will be given in square brackets, '"Payera [Danais]
e.g.
&
hakeriana^^ bakeriana (Homolle) Buchner Puff Danais bakeriana Homolle). Coursiana
for P. (syn. will
be shown to be congeneric with Payera (see Appendix) to stress the original generic association of
its
;
only species, "\Payera [Coursiana] homolleana'' will occasionally be used.
TAXONOMIC
HISTORY
(Table
III. 1)
Homolle
M""^
Mala
New
Baron and of French such as Perkier de la Bathie, Decary, or Humbert.
collectors
taxa v^ere added by Cavaco (1965, 1966 and 1968Z?). In one of his publications (Cavaco,
Comoro
1966), he included a key to the Malagasy and Island species he recognized.
Unfortunately, publication did not give any details on the species and, moreover, also
this
included some new names which were not formally validated.
The same with Schismatoclada, a genus endemic to the East Malagasy
authors also dealt
region (Homolle, 1939; Cavaco, 1964, 1967, 1968*7). Both of them accepted the genus as
it
was and did not discuss the relationships to Danais.
at
all
(Robbrecht, Danais and
most treatment of the Rubiaceae 1988),
In the recent overall
whose taxonomic
amongst genera position uncertain. In the
Schismatoclada are hsted those is
past, the two genera have on the one hand been associated with the tribe Cinchoneae (subfam.
Cinchonoideae) and, on the other, been referred to tribe Hedyotideae (subfam. Rubioideae) (see
&
Andersson
Persson
Table Dealing with the circumscription of the Cinchoneae, (1991)
1).
;
26
provisionally transferred both genera to the Hedyotideae. The arguments that have, over the
two
placement of of the
been brought forward favour of or against a particular either
years, in
genera will be discussed in the final chapter.
T
Table Tribal position of Danais, Schismatoclada, Coursiana and Payera (year of esta-
—
1
:
Abbreviation of Cin
blishment brackets) according to different authors. tribes
in
: :
Hed Cin/Hed Cin-C Cinchoneae-Cin-
Cinchoneae; Hedyotideae; intermediate;
: ;
:
choninae not dealt with.
Danais Schismatoclada Coursiana Payera
Author
(1799) (1883) (1942) (1878)
Hooker Drake
Endlicher
(1838), (1873),
Bremekamp
Cin
(1948)
(1898Z>),
"
"
Baillon Cin Genipeae 1
(1880)
Baker Homolle Cin
(1883), (1939)
"2
"
Schumann
Cin Cin Oldenlandieae
(1891)
Homolle Cin
(1942)
BOITEAU Cin Cin
(1941)
Bremekamp Hed Hed
3
(1952, 1966) [Ci n]
Verdcourt Hed Hed Hed
(1958)
Cavaco
Cin Cin Cin
(1964, 1966, 1968a, 1968Zj)
Robbrecht Cin/Hed Cin/Hed Cin-C Hed
(1988)
&
Andersson Persson Hed Hed* Hed*
(1991)
now
Myrioneuron" Robbrecht,
"nr. placed
1. (the latter is in tribe Isertieae; cf. 1988).
± Hed
2. Is see text.
;
Not specifically mentioned; infers that either the inclusion in Cin was not doubted or that the genus was
3.
unknown
him;
to see text.
Provisional transfer suggested.
4.
The genus Coursiana was established 1942 by M""^ Homolle, although she never
in
specifically mentioned a type species. This was eventually rectified by Cavaco (1968a) by
describing Coursiana homoUeana, the only species ever described in this East Malagasy genus.
'^ Homolle (1942) had made absolutely no reference to Danais and Schismatoclada when
she described Coursiana, although she had dealt with the two genera in 1936 and 1939,
respectively (see above). It difficult to understand that she should not have been aware of the
is
Cavaco
between
close association these three genera. (1968a) did not discuss the relationships
and Homolle's
of Coursiana either uncritically accepted M""^ suggestion of placing the genus
in the tribe Cinchoneae (subtribe Cinchoninae).
The position of Coursiana was also accepted by Robbrecht although both
tribal (1988),
Bremekamp and Verdcourt had
(1952, 1966) (1958) suggested a transfer to the Hedyotideae,
&
Andersson Persson (1991) also favoured a position in the latter tribe.
P
Notes on and attached to herbarium sheets in indicate that Prof. J.-F. Leroy, former
director of the Paris herbarium, dealt with Danais, Schismatoclada and Coursiana 1975.
in It
who
was he discovered that the rather ill-known monotypic Malagasy endemic genus Payera
is
27
apparently closely allied to the above mentioned complex of genera. can be reconstructed
It
from his notes that he initially intended to transfer Coursiana and species of SchismatocJada to
He
Payera. later changed his mind and apparently thought that the complex of four genera
is
best combined into one genus, Danais Ventenat [^ Payera Baillon
(1799) (1878),
i.e.,
= =
Baker Homolle
Schismatoclada Coursiana Prof. Leroy, however, never
(1883), (1942)].
published
his findings.
"
While Prof. Leroy's concept of ''Danais not accepted below), discovery of
(see his
s.l. is
the genus Payera being part of this alliance certainly a highly valuable contribution which,
is
unknown
most would have remained
otherwise.
likely,
The "
genus Payera^ originally placed in the Genipeae'' (Baillon, 1880), was considered
member
some
to be a of the Oldenlandieae (which to extent correspond to the Hedyotideae as
Schumann
circumscribed today) by (1891). The ill-known genus was not specifically dealt with
Robbrecht
by and
other authors, (1988) accepted position in the Hedyotideae,
its
ri t
THE CHARACTER
RESULTS STATES
IV.
:
VEGETATIVE CHARACTERISTICS
Growth
form, stems
1.
woody
Danais The genus comprised of lianas or climbing shrubs, some of which
entire
is
:
may reach considerable dimensions. D, magna and D. ligustrifolia^ for example, are reported to
m
canopy
produce shoots which reach ca. 25 into the of rain forest trees. Others are less
extensive; shoots apparently do not become longer than a few meters.
their
many
stemmed
The lianas of Danais appear to be several - to - throughout. According to
much-branched
observations by one of us (C, stems are normally rather above.
field P.),
Ultimate branches do not twine.
" "
Descriptions of climbing shrubs indicate that the basal parts of plants indeed have a
many
stemmed), from which markedly
fundamentally shrubby structure several to
- -
(i.e.
elongated and climbing axes
arise.
If the main shoots are not yet so long that they can climb up on the surrounding higher
may "
vegetation (younger individuals plants resemble typical " shrubs. Moreover, plants
!),
downwards
which bend
growing surroundings which lack higher vegetation, develop shoots
in
+
ground and even creeping shoots has, for example, been noted for D. coronata,
to the (this,
on herbarium of Danais
which grows edge vegetation). Descriptions labels
typically in forest
and somewhat
misleading
species as shrubs probably refer to such situations are, therefore, in
that they do not the true situation. —
reflect
somewhat
In Danais, younger shoots are either glabrous or hairy. If present, hairs
—
and throughout are normally rather short; only in a few species
uniseriate several-celled
(e.g. D. vestita and D. hispida), relatively long hairs and a dense stem indumentum are present.
become
Older stems tend to glabrescent.
28
The
The lianaceous habit separates Danais from Schismatoclada and Payera, latter are
up few meters
normally described as either shrubs or small trees to a tall.
In Schismatoclada, shoots are typically glabrous, while in Payera they are often densely
youngest
hairy (at least the parts).
show
conspicuous sympodial-dichasial
In both, the ultimate parts of twigs often
branching (because of the invariably terminal position of the inflorescences).
and
Leaves Stipules
2.
Stipules
a.
and
In genera dealt with, the interpetiolar stipules are variable in size shape.
all
+ + r mm,
rounded
Danais typically has triangular to deltoid or stipules (up to ca. 8 rarely
mm
to 15 long). There are few exceptions, e.g. D. andribensis with consistently bifid stipules
many
and D. a with stipules which bear several to fimbriate appendages; D.
vest
it
brickavillensis has rounded stipules with laciniate margins.
and
In both Schismatoclada Payera, the stipules are either entire or variously divided.
Quite frequently, the stipules of Schismatoclada are
mm
small (to ca. 3 long) and often entire (e.g. 5. viburnoides. Fig. 4, G) but there are also
Cavaco,
species with rather large, pluri- to multifimbriate stipules 5. longistipula,
(e.g. cf.
1967 Fig. 7-2). In Payera, a very clear trend towards large, conspicuous, almost leaf-like
1,
:
stipules is noticeable (e.g. P. glabrifolia. Fig. 4, E). P. decaryi has elliptic to lanceolate-elliptic
mm
Not uncommonly
stipules which are to ca. 45 long. the stipule margins are laciniate or
bear fimbriate appendages. Within the genus there are, however, also small-stipuled species
mm
P, madagascariensis, stipules only a few long).
(e.g.
+
many
In taxa of three genera, the stipules are persistent and frequently recognizable
all
as dried, somewhat hardened structures on old shoot portions. Large-stipuled Payera species
appear to be an exception (stipules often disappear together with the leaf-blades).
CoUeters are normally present on the inside and on the margins of the stipules. The
known many
conical colleters correspond to the standard type which to occur in other taxa
is
Robbrecht,
of Rubiaccae
the 1988).
(cf.
morphology
Leaf
b.
In genera dealt with, decussate leaf arrangement predominates. In a few taxa of
all
Danais, however, true whorls of three and/or four do occur regularly D. coronata, D.
(e.g.
xerticillata and volubilis; within D, cernua, both decussate and ternate leaf arrangement was
Z>.
observed). Also Schismatoclada marojejyensis has leaves arranged whorls of
in three.
Damns
The leaves are typically petiolate. In and Payera, petiole lengths range from
ca. 1.5
mm.
some morphs
to ca. 35 In taxa, e.g. of /). pubescens or P. coriacea, petioles are short
mm
mm); The
(ca. 1.5-3 D. verticillata has subsessile leaves. longest petioles, to ca. 30-35 long,
magna and
occur in Z). /). decaryi. In Schismatoclada, petioles tend to be very short to short
mm); may
to 10 due to long-cuneate leaf-blade bases leaves even be subsessile S,
(i.e., (e.g.
I
humbertiana).
29
and
Sizes shapes of the leaf-blades vary considerably within the genera (and sometimes
also within individual
species).
Common
mm.
from
are size ranges ca. 20 to 120 Exceptions are P. decaryi, with
leaf-
mm
&
morphs
blades to over 300 long, of D. pubescens nummular Puff
("Z). olia-torms"', see
if
mm
BuCHNER,
in press) with blades less than 20 long, and S. marojejy The latter a
crisis. is
10mm
microphyllous more
species (blades hardly than long) occurring in ''cricoid" high
mountain
vegetation.
Obovate,
ovate, elliptic, elliptic-lanceolate to hnear-lanceolate leaf-blade shapes are
frequent. Apices are often acuminate (very conspicuous, long-acuminate D.
in, e.g.,
+
tsaratauanensis), shortly pointed or rounded; emarginate apices morphs of D.
(e.g. in
pubescens) are rather rare. The bases of the blades are most commonly + cuneate.
( )
Most
(or all?) taxa of all three genera appear to have evergreen leaves. While the leaves of
all Schismatoclada species seem to be invariably rather thick and leathery, both in Danais and
Payera, two groups of taxa can be distinguished. One has thick, leathery (distinctly coriaceous)
The
leaves (e.g. D.fragrans, D. vohibilis, D. coronata; P. coriacea). other has leaves which tend
be
to relatively thin (less distinctly coriaceous; e.g. /). microcarpa, D. ligustrifolia; P, decaryi).
The
rather thinnish leaves of some taxa of Payera might be short-lived and deciduous, but
this
needs confirmation
in the
field.
Especially for taxa of Payera seems to be characteristic that leaves are crowded on the
it
very terminal parts of long shoots P. decaryi. Fig. D).
(e.g. 4,
In Danais and Payera^ there are both glabrous and hairy taxa; in contrast, the leaves of
Schismatoclada tend to lack an indumentum. In hairy taxa, the indumentum, although
invariably comprised of uniseriate, pluricellular hairs, varies in extent and length and shape of
the hairs. The leaf blades are either hairy above and below P. homolleana), or only on one
(e.g.
may
of the surfaces, or only on the midrib below P. beondrokensis). Hairs be short or long
(e.g.
+
and curled and dense (woolly to velvety indumentum, D. D. pubescens).
e.g. vestita,
indumentum
Rather frequently, the varies within taxa, in essentially "hairy'' taxa,
i.e.,
there may be morphs which the indumentum only sparsely developed or almost lacking
in is
(e.g. in D. hispida or D. pubescens).
The venation of the leaf blades conspicuous in some taxa of Danais and Payera. Due to
is
may
raised secondary and higher order veins, reticulate venation patterns either be very
prominent on the upper (e.g. D. aurantiaca) or on the lower blade surface (e.g. P. homolleana).
Leaf anatomy Table
(Fig. 1-2; 2)
c.
—
hypodermis
epidermis;
Epidermis, incompletely multiple
In three genera dealt with, the lower epidermis is always smaller-celled than the upper.
all
and stomata of
Without was found be one-layered contains the (leaves taxa
exception, to all
it
are hypostomatic). The stomatal apparatus invariably of the characteristic "rubiaceous
is
The sometimes F; Very
type" (Wilkinson, stomata are slightly raised Fig. F).
1979). (cf. 1, 2,
markedly and conspicuously raised stomata, as depicted for P. [Coursiana] homolleana
Wax
(HoMOLLE, were not detected. crystals were rather frequently noted in the
1942; Fig.
2),
B-Q.
area of the stomata and also elsewhere on the lower surface (e.g. D. fragrans. Fig.
1,
30
PP
V
t
^-™^
^^^-fc-
£ rJ
_
Sr- J
-^- -^
'>^-^r^'
M. -^-
A2
-^
n
^
"p
^
:>
*
^A-^
^ 1^
Fig. 1. A, D-H, leaf sections of Danais species A, D.fragram; D-F, D. voluhilis; E, detail showing incompletely
:
multiple epidermis and palisade cells; F, detail of spongy mesophyll, lower epidermis and stoma; G-H, with
species
—
epi- and hypodermis G, D. pubescens —H, D. microcarpa, B-C, SEM-graphs of lower leaf surface of D,fragrans
; ;
showing wax and stoma
crystals (C). ep, epidermis; hd, hypodermis; palisade parenchyma; spongy
— pp, sp,
^m = D - G = ^
mesophyll. Scale bars 100 (A H; B; E F); 10 am
(C).
:
31
As
regards the upper epidermis, the situation heterogeneous. In Schismatoclada and
is all
but one of investigated Payera species, there a single layer of large cells (Fig. C-D). In
is 2,
Danais, there are two groups of taxa. In one (taxa with relatively thin, although evergreen
D, microcarpa and Table ''normal" upper
leaves, others, a single-layered epidermis
cf.
2), is
present, but below there a conspicuous, continuous hypodermis comprised of
layer
it is cells
G-H
roughly three times as large as those of the epidermis (Fig. A-B). In the other group
2,
1,
,
(taxa with thick, leathery leaves, D. fragrans and others, Table there an
cf. 2), is
"incompletely"
not continuously) several-(normally two-)layered epidermis A,
(Fig.
(i.e., 1,
As
D-E), opposed to a true hypodermis (for information on the ontogeny of leaf blades and
of hypodermal Napp-Zinn,
the origin layers see 1973), the multiple epidermis apparently
comes about by the development of cross walls parallel to the upper-surface. This
characteristic epidermis structure has, by the way, been illustrated correctly for D. volubilis by
&
DuBARD Dop
(1925 Fig. IV, as Alleizettea bracteata'').
7,
''
:
may
The
upper
epidermis has a cuticle which, particularly in the coriaceous-leaved taxa,
be very thick and conspicuous. The cuticle of the lower epidermis generally considerably
is
thinner.
and
Table anatomical characters of Danais, Schismatoclada Payera.
2 Selected leaf
:
herbarium mate-
on of preserved material others
based the investigation (all
:
hypodermis
several layers; see text.
rial).
tf
Rather upper Large-celled, mostly incomple- ''Normal" large-celled upper epi-
small-celled epi-
hypodermis multiple upper epidermis; dermis; hypodermis absent
dermis, large-celled tely
hypodermis absent
below
it
Schismatoclada
Danais
Danais
aiirantiaca :
:
humbertiana
D. cernua D. andribensis S.
marojejyensis
D. D. coronata S.
hispida
C)
D. A-B) D, dauphinensis S. psychotrioides (Fig. 2,
humblotii'' (Fig.
2,
* A)
D, fragrans (Fig
D.
ligustrifolia 1,
D. magna D. terminalis
H) D. tsaratananensis
D. microcarpa"*^ (Fig.
1,
D.
D. nigra vestita
D-F)
D. pubescens"^ (Fig. G) D. volubilis'' (Fig. 1,
1,
D.
rhamnifolia
D.
verticillata
Payera
:
D)
P. decaryi (Fig.
E-F)^ 2,
Payera
coriacea* (Fig.
2,
homoUeana
P. [Coursiana]
—
Mesophyll
much
parenchyma always extensive than the spongy
paUsade less
In three genera, the is
all
The normally made up of two to three layers of elongated, cylindrical
mesophyll. former
is
+
D. D-E) to isodiametric (e.g. P. decaryi. Fig. 2, Z)), rather tightly
Fig.
(e.g. volubilis. 1,
Numerous with raphides tend to present particularly in the
packed
idioblasts filled
cells.
32
J- •'^
—
Fig. 2. Leaf sec—tions A-B, Danais f—aumblotii; B, detail showing epi- and hypodermis. C, Schismatoclada
:
psychotrioides. D, Payera decaryi. E-F, P. coriacea; E/note epidermis and several hypodermis layers an—d
vascular bundles with sclerenchymatic bundle sheaths; F» part of spongy mesophyll, lower epidermis and stoma,
— —
D =
=
B
C, E-F, from reconstituted herbarium material. Abbreviations as in Fig. Scale bars 100 ^zm (A
1.
: ;
F;C=E).