Morphology of cypselas in subtribe Arctotidinae (Compositae– Arctotideae) and its taxonomic implications

Abstract

Subtribe Arctotidinae comprises an estimated 80-90 species classified in five genera. Taxonomic relationships within the subtribe are poorly understood and the taxonomy of Arctotis is notably confused. Cypselar morphology provides a suite of taxonomic characters important at the genus and species levels in Arctotidinae. To help resolve taxonomic relationships in the subtribe, a scanning electron microscopic survey of the cypselae of 40 species from four genera (Arctotheca, Arctotis, Cymbonotus, and Haplocarpha) was undertaken. Thirteen informal species groups are distinguished based on cypselar morphology. The results indicate that a reappraisal of the generic concepts of Arctotis, Cymbonotus, and Haplocarpha is warranted. A close affinity is implicated between Arctotis sect. Austro-orientales and Cymbonotus. Arctotis species are assigned to nine species groups. Lewin's infrageneric classification of Arctotis is partially supported. Haplocarpha comprises three disparate species groups. Arctotheca forms a morphologically coherent group based on cypselar morphology herein. A diagnostic key to these informal groups based on cypselar morphology is presented.

Full text

A
NN
.M
ISSOURI
B
OT
.G
ARD
. 92: 569–594. P
UBLISHED ON
28 D
ECEMBER
2005.
MORPHOLOGY OF CYPSELAE
IN SUBTRIBE ARCTOTIDINAE
(COMPOSITAE–
ARCTOTIDEAE) AND ITS
TAXONOMIC IMPLICATIONS
1
R. J. McKenzie,
2
* J. Samuel,
2
E. M. Muller,
2
A. K. W. Skinner,
2
and N. P. Barker
2
A
BSTRACT
Subtribe Arctotidinae comprises an estimated 80–90 species classified in five genera. Taxonomic relationships within
the subtribe are poorly understood and the taxonomy of Arctotis is notably confused. Cypselar morphology provides a
suite of taxonomic characters important at the genus and species levels in Arctotidinae. To help resolve taxonomic
relationships in the subtribe, a scanning electron microscopic survey of the cypselae of 40 species from four genera
(Arctotheca, Arctotis, Cymbonotus, and Haplocarpha) was undertaken. Thirteen informal species groups are distinguished
based on cypselar morphology. The results indicate that a reappraisal of the generic concepts of Arctotis, Cymbonotus,
and Haplocarpha is warranted. A close affinity is implicated between Arctotis sect. Austro-orientales and Cymbonotus.
Arctotis species are assigned to nine species groups. Lewin’s infrageneric classification of Arctotis is partially supported.
Haplocarpha comprises three disparate species groups. Arctotheca forms a morphologically coherent group based on
cypselar morphology herein. A diagnostic key to these informal groups based on cypselar morphology is presented.
Key words: Arctotheca, Arctotis, Asteraceae, Cymbonotus, fruit morphology, Haplocarpha, scanning electron micros-
copy, systematics.
Arctotideae comprises about 220 species with a
principally southern African distribution. It in-
cludes well-known horticultural subjects such as
Arctotis L. and Gazania Gaertn. Two subtribes are
currently recognized: Arctotidinae and Gorteriinae
(Bremer, 1994). Previous workers have recognized
other subtribes such as Gundeliinae (Bentham,
1873a, b, as Gundelieae) and Eremothamninae
(Leins, 1970). Molecular studies indicate the Gun-
deliinae is sister to the Cichorieae Lam. & DC.
(Lactuceae Cass.) (Karis et al., 2001; Funk et al.,
2004). Recent molecular evidence places Eremo-
thamnus O. Hoffm., Heterolepis Cass., and Hoplo-
phyllum DC. with Arctotideae, but outside Arcto-
tidinae and Gorteriinae (Funk et al., 2004).
However, the affinities of these genera indicated by
morphological evidence are much less certain (Bre-
mer, 1994; Funk et al., 2004).
A study of plastid and nuclear DNA sequences
in Arctotideae provided support for the monophyly
of Arctotidinae (Funk et al., 2004). However, some
phylogenetically important species of Arctotidinae,
such as any representative of Haplocarpha subg.
1
We gratefully acknowledge the Rhodes University Joint Research Committee and the National Research Foundation
of South Africa (grant numbers 2042600, 2046932, and 2053645 to NPB, and a Postdoctoral Fellowship to RJMcK)
for financial support; the curators of the BOL, ETH, GRA, NBG, NH, NSW, and PRE herbaria for the loan of, or access
to, specimens; Estelle Brink, John Pruski, John Strother, and the editor for invaluable suggestions to improve the
manuscript; Shirley Pinchuck and Robin Cross of the Rhodes University Electron Microscope Unit for technical assis-
tance with electron microscopy.
2
Department of Botany, Rhodes University, P.O. Box 94, Grahamstown, 6140, South Africa. r.mckenzie@ru.ac.za;
n.barker@ru.ac.za.
Landtia (Less.) Beauverd, were not included in
their study. Furthermore, identification of morpho-
logical synapomorphies for the subtribe remains
more difficult. Funk et al. (2004) listed ten mor-
phological traits characterizing Arctotidinae, but
noted that seven of those would be considered ple-
siomorphic, and the absence of latex is probably
apomorphic but difficult to utilize at the subtribal
level in the subfamily Cichorioideae. This leaves
two possible synapomorphies for Arctotidinae: a
swelling on the disc-floret style with a ring of
sweeping papillae below the apical branches, and
3 to 5 well-developed ridges or wings on the fruits.
As now circumscribed, Arctotidinae comprises
approximately 80 to 90 species in five genera (Nor-
lindh, 1977; Bremer, 1994). Most of the species (an
estimated 60 to 70 species) belong to Arctotis,
which is currently under revision in our laboratory.
Of the remaining genera, Arctotheca J. C. Wendl.
contains five species, including the recently de-
scribed A. marginata Beyers (Beyers, 2002). Hap-
locarpha Less. includes nine species, Dymondia
Compton is monotypic (Compton, 1953; Rourke,

570 Annals of the
Missouri Botanical Garden
1974), and the Australian endemic Cymbonotus
Cass. has two species. A third Australian-endemic
species, Arctotis maidenii, was treated as C. sp. ‘A’
by Murray (1992). An additional four genera (Cryp-
tostemma R. Br., Landtia Less., Microstephium
Less., and Venidium Less.) recognized by Harvey
(1865) are now included within other genera: Land-
tia was merged with Haplocarpha, Venidium with
Arctotis (Beauverd, 1915), and Cryptostemma and
Microstephium with Arctotheca (Lewin, 1922a).
A cladistic analysis of morphological characters
by Bremer (1994) showed the relationships within
Arctotidinae as an unresolved polychotomy. More
recent molecular studies (Karis et al., 2001; Funk
et al., 2004) have not included sufficient samples
to address generic-level relationships, focusing in-
stead on tribal and subtribal relationships. Ongoing
morphological and molecular studies in our labo-
ratory indicate generic concepts in Arctotidinae re-
quire further resolution. Delimitation of genera is a
frequent problem in Compositae (Cronquist, 1985),
often reflecting the occurrence of morphologically
intermediate taxa or varied combinations of taxo-
nomically important characteristics across a group.
Instability in generic classification in Arctotidinae
partly reflects differences in the characters empha-
sized for delimiting genera by different authors. For
example, Lessing (1832) emphasized fruit and pap-
pus morphology, ray floret fertility, and filament or-
namentation. However, fruit pubescence and pap-
pus characters were found to be unreliable generic
criteria in the group and floret fertility received
greater emphasis (Beauverd, 1915; Lewin, 1922a).
In addition, considerable confusion exists re-
garding the taxonomy of Arctotis. This partly reflects
the brevity of most early species descriptions and
a lack of detail on characters now considered to be
taxonomically important. Resolution of this confu-
sion will be best achieved by a careful reappraisal
of the taxonomic utility of previously utilized char-
acters, a search for additional taxonomically useful
characters, and collection of additional specimens
of poorly known species. Owing to the previous em-
phasis placed on fruit morphology (e.g., Lessing,
1832; Harvey, 1865; Beauverd, 1915; Lewin,
1922a), we present here an investigation of the tax-
onomic utility of fruit morphology in Arctotidinae.
Examination of other morphological features and
phylogenetic analysis of DNA sequence data are
ongoing in our laboratory.
Compositae fruits are termed cypselae. These are
indehiscent, one seeded, usually dry, derived from
inferior, bicarpellate ovaries, and the testa is adnate
to the endocarp (Bremer, 1994). Cypselar morphol-
ogy was emphasized for classification of Arctotidi-
nae by Lessing (1832), who distinguished eight
genera based on cypselar and pappus morphology,
ray-floret fertility, and filament ornamentation. Can-
dolle (1837), Harvey (1865), and Bentham (1873a,
b) essentially followed Lessing’s generic classifica-
tion. Beauverd (1915) and Lewin (1922a) gave less
weight to cypselar pubescence and pappus mor-
phology and greater emphasis to ray-floret fertility
to discriminate genera. Both authors used cypselar
morphology and filament ornamentation to distin-
guish infrageneric taxa in Arctotis and Haplocarpha.
Beauverd (1915) recognized four subgenera in Arc-
totis based on cypselar and pappus morphology and
distinguished two sections within subgenus Arctotis
based on growth form. Landtia was recognized at
subgeneric level within Haplocarpha based on pap-
pus characters. Lewin (1922a) recognized 15 sec-
tions within Arctotis differentiated on the basis of
morphology of the cypselae, involucral bracts, and
leaves. These sections did not follow previously
recognized species groupings. The species formerly
segregated as Venidium were placed in five sec-
tions, and one of those sections (sect. Acuminatae
K. Lewin) includes a mixture of species that were
placed previously in Venidium and Arctotis.
In Arctotidinae, the cypsela pericarp is very hard
and contains a sclerenchymatous layer (Reese,
1989). A diagnostic feature of cypselae in Arctotis
is the presence of two or three well-developed
wings on the abaxial surface; the lateral wings are
curved inward creating either one or two furrows or
chambers depending on the extent of development
of the median wing. Often these furrows or cham-
bers have been termed ‘‘cavities’’ (e.g., Norlindh,
1977; Goldblatt & Manning, 2000), and to avoid
confusion this term is retained herein. Cypselar
wings are weakly developed in Cymbonotus and ab-
sent in Arctotheca and Haplocarpha (Beauverd,
1915; Lewin, 1922a). Instead, cypselae in Arcto-
theca and Haplocarpha usually have three to five
longitudinal ridges, which are more strongly devel-
oped on the abaxial surface. Based on cypsela anat-
omy, especially pericarp formation and the number
of vascular traces, the cypselae of H. rueppellii and
H. scaposa are distinct from those of Arctotheca,
Arctotis, Cymbonotus, and H. lyrata (Reese 1989).
Herman (2001) observed and illustrated three tri-
chome types (multicellular biseriate, septate, and
twin hairs) present on the cypselae of Arctotheca,
Arctotis, and Haplocarpha species. Twin hairs are a
distinctive trichome type widespread in Compositae
(Hess, 1938). Some species possess a basal ring of
trichomes attached to the carpopodium; such a ring
of trichomes is termed a coma by some authors
(Hilliard, 1977; Herman, 2001). The pappus, when

Volume 92, Number 4
2005
571McKenzie et al.
Morphology of Cypselae
present, comprises one or two whorls of free scales
in most species; when two whorls are present, the
scales are alternately arranged. In some species,
the pappus is uniseriate and the scales are at least
partially fused at the base, forming a terminal
crown (herein termed coronate).
In this paper, cypselae of representative species
from each of the genera in Arctotidinae (with the
exception of Dymondia) are examined by scanning
electron microscopy (SEM) with the purpose of
identifying ‘‘natural’’ species groups in Arctotidinae
based on similarities of external fruit morphology.
These groupings will help provide a sound basis for
resolution of generic concepts in conjunction with
evidence from other sources.
M
ATERIALS AND
M
ETHODS
Cypselae containing a mature seed were ob-
tained from herbarium specimens and plants in the
wild. Cypselae from 4 species of Arctotheca, 25 spe-
cies of Arctotis comprising at least one representa-
tive from each of Lewin’s (1922a) 15 sections, 2
Cymbonotus species, and 9 Haplocarpha species
were examined (Table 1, Appendix 2). Cypselae
containing a mature seed were unavailable for
study for Dymondia margaretae Compton. The
South African species recognized generally follow
Germishuizen and Meyer (2003). Goldblatt and
Manning (2000) treated A. perfoliata as a synonym
of A. discolor (Less.) Beauverd, but the name A.
perfoliata has nomenclatural priority for the com-
bined taxon.
Cypselae were examined by means of a JEOL
JSM 840 Scanning Electron Microscope at the
Rhodes University Electron Microscope Unit. All
samples were dry and thus were not further treated
before being sputter-coated with gold-palladium. In
some species the cypselar surface has a dense cov-
ering of trichomes or a basal coma, which, in some
instances, were removed to allow a clearer vision
of the cypselar surface. Supplementary observa-
tions of trichome structure were made with a Wild
compound microscope.
Cypselar ontogeny was not investigated. It is im-
portant to note that cypsela shape, size, and surface
features may be influenced by the proportion within
a capitulum of cypselae containing a seed and the
stage of cypsela development (Breitwieser & Ward,
2003). An additional factor to consider is homology.
The ray florets in Arctotheca and Arctotis sect. An-
omalae K. Lewin, and the central disc florets in
most species of Arctotis, are female sterile (Lewin,
1922a). In A. fastuosa, the ray and disc florets pro-
duce cypselae with minor morphological differenc-
es (Beneke et al., 1992). Where available, cypselae
containing a mature seed from both ray and disc
florets were examined. Because only minor differ-
ences in cypselar morphology were noted (princi-
pally dimensions and pubescence), separate de-
scriptions of cypselae from both floret types are not
presented. Absolute quantitative characters were
excluded; two ratios (maximum pappus length: cyp-
sela length and maximum coma length: cypsela
length) were included.
R
ESULTS
SEM micrographs of species examined are shown
in Figures 1–7. Selected cypselar characters for all
species are summarized in Table 2. Thirteen infor-
mal species groups (I–XIII) are recognized herein
on the basis of fruit morphology.
K
EY TO THE
S
PECIES
G
ROUPS IN
A
RCTOTIDINAE
1. Prominently raised ridges or wings absent on
both abaxial and adaxial surfaces
-------------
Group I
1
9
. Abaxial surface with prominently raised ridges
or wings
-----------------------------------------------------------------
2
2. Abaxial surface with 2 or 3 prominently raised
longitudinal ridges
------------------------------------------------
3
2
9
. Abaxial surface with 2 or 3 distinct wings
---------
5
3. Adaxial surface smooth; pappus well developed,
length at least 1.5
3
cypsela
-------------------
Group II
3
9
. Adaxial surface transversely rugose; pappus at
most equal in length to cypsela
---------------------------
4
4. Ridges strongly dentate and rugose; uniseriate
trichomes present; coma absent
--------------
Group III
4
9
. Ridges not dentate or rugose; clothed with sep-
tate trichomes often forming woolly indumentum
completely concealing cypsela and pappus;
coma of either twin hairs or multicellular bis-
eriate trichomes present, neither very distinct
from septate trichomes
----------------------------
Group IV
5. Wings developed on both abaxial and adaxial
surfaces
------------------------------------------------------------------
6
5
9
. Wings developed on abaxial surface only
-----------
7
6. Lateral abaxial wings wide-spreading, median
abaxial wing absent, resulting in a single often
broad abaxial cavity; adaxial wings very re-
duced
----------------------------------------------------------
Group V
6
9
. Lateral abaxial wings strongly incurved, median
abaxial wing
6
concealed and the two abaxial
cavities occluded; adaxial wings well developed,
forming two narrow lateral cavities
---------
Group VI
7. Cypsela with a single abaxial cavity; median ab-
axial wing weakly developed
-----------------
Group VII
7
9
. Cypsela with two abaxial cavities; in some spec-
imens cavities almost occluded by the strongly
incurved abaxial wings
-----------------------------------------
8
8. Pappus absent
---------------------------------------
Group VIII
8
9
. Pappus present
------------------------------------------------------
9
9. Cypsela apical plate offset to abaxial side with
strongly convex adaxial surface; trichomes sim-
ple, uniseriate; pappus scales to 0.25 mm long;
coma absent
---------------------------------------------
Group IX
9
9
. Cypsela apical plate not offset to abaxial side;
septate trichomes present; maximum pappus

572 Annals of the
Missouri Botanical Garden
scale length exceeding 0.25 mm long; coma of
twin hairs or glandular biseriate trichomes pre-
sent on carpopodium
------------------------------------------
10
10.
Cypsela wings minutely dentate; sparse septate
trichomes to 135
m
m long; coma of partial or
complete ring of glandular biseriate trichomes
on carpopodium; pappus very reduced, with in-
ner whorl of 5 to 7 scales, adaxial scales usually
absent, sometimes with rudimentary scales of
outer whorl on abaxial side
-----------------------
Group X
10
9
.
Cypsela with entire or dentate wings; glabrous
or with septate trichomes exceeding 135
m
m
long; coma of twin hairs present; pappus in 2
whorls, inner whorl much longer than outer
whorl
--------------------------------------------------------------------
11
11.
Cypsela wings fused to median abaxial rib in
proximal part; cypsela base constricted; abaxial
cavities ovate
--------------------------------------------
Group XI
11
9
.
Cypsela wings not fused to median abaxial rib
in proximal part; cypsela base not constricted;
abaxial cavites obconical, obovate, or linear
----
12
12.
Lengths of pappus scales and coma less than
half the cypsela
---------------------------------------
Group XII
12
9
.
Lengths of pappus scales and coma subequal or
longer than cypsela
-------------------------------
Group XIII
SPECIES GROUP I
Haplocarpha hastata (Fig. 1A), H. nervosa (Fig.
1B), H. rueppellii (Fig. 1C)
The species group comprises three species as-
signable to Haplocarpha subg. Landtia. The cyp-
selae have a smooth surface, a very reduced pap-
pus, and lack prominent abaxial ridges or wings.
The cypselae are obovoid in H. nervosa and ob-
ovoid-elliptic in H. hastata and H. rueppellii; the
cypsela base is constricted in H. rueppellii. The
cypselae are tangentially flattened in H. hastata,
but in H. nervosa and H. rueppellii the cypselae are
somewhat 3- or 4-angled; in each species the ad-
axial surface is more strongly convex than the ab-
axial surface. The cypselar surface is smooth in H.
hastata and H. rueppellii; in some specimens of H.
nervosa the surface is transversely rugose. The cyp-
selae are glabrous in H. hastata and H. rueppellii;
either biseriate glandular trichomes or uniseriate
trichomes to 125
m
m long are present on cypselae
in H. nervosa. A coma of twin hairs to 1 mm long
is present in H. nervosa and is lacking in H. hastata
and H. rueppellii. The pappus is coronate with in-
distinct lobes in H. hastata. Haplocarpha nervosa
and H. rueppellii have a pappus of scales in 2
whorls, the scales of each whorl alternate with those
of the other whorl. In H. nervosa, the inner whorl
(0.3–0.5 mm long) is slightly longer than the outer
whorl (0.2–0.3 mm long); in H. rueppellii the scales
are of uniform lengths.
SPECIES GROUP II
Haplocarpha lanata (Fig. 1E, F), H. lyrata (Fig.
1D), H. oocephala (Fig. 1G), H. parvifolia, H.
scaposa (Fig. 1H)
The species group comprises four species placed
in Haplocarpha subg. Haplocarpha by Beauverd
(1915) and H. oocephala, which was recently trans-
ferred to Haplocarpha from Arctotis (Beyers 2000).
The cypselae are broadly obconical in H. scaposa
and obconical in the other species, tangentially flat-
tened in H. lanata, H. lyrata, and H. scaposa, with
well-developed abaxial ridges, a coma of twin hairs,
and a pappus of scales in 1 or 2 whorls, which is
always longer than the cypsela. The ridges are least
developed in H. scaposa. In H. lyrata the lateral
abaxial ridges are shortly incurved toward the me-
dian ridge (most evident in transverse section). The
ridges are very broad and separated by narrow fur-
rows in H. oocephala and H. parvifolia. The ridges
are transversely rugose in H. oocephala and smooth
in the other species. The adaxial surface is smooth
in all species. The cypsela surface is glabrous in
H. scaposa. Dense twin hairs are present on both
abaxial and adaxial surfaces in H. lanata and H.
lyrata. Glandular biseriate trichomes are present on
the abaxial ridges in H. oocephala and H. parvifol-
ia. All species possess a coma of twin hairs; the
coma is subequal to the cypsela in H. lanata and
H. lyrata, up to double the cypsela length in H.
oocephala and H. parvifolia, and more than triple
the cypsela length in H. scaposa. There are 2 pap-
pus whorls present in H. oocephala, H. parvifolia,
and H. scaposa, of which the inner whorl is much
longer than the outer whorl; only the inner pappus
whorl (as indicated by the alignment of the scales
with the cypsela ridges) is present in H. lanata and
H. lyrata. The pappus scale apices are acuminate
in H. lyrata and H. scaposa, acute in H. lanata,
and obtuse to rounded in H. oocephala and H. par-
vifolia. The scale apex is serrate in H. lanata and
H. lyrata. The scales are more than 4 times the
cypsela length in H. scaposa and 1.5 to 2 times the
cypsela length in the other species.
SPECIES GROUP III
Haplocarpha schimperi (Fig. 2A–C)
On the basis of external cypselar morphology,
Haplocarpha schimperi is anomalous within H.
subg. Landtia. The cypselae are obovoid, tangen-
tially flattened with 3 well-developed, dentate ab-
axial ridges, transversely rugose on both abaxial
and adaxial surfaces, with uniseriate trichomes to
75
m
m long and a short pappus. A coma is absent.

Volume 92, Number 4
2005
573McKenzie et al.
Morphology of Cypselae
Table 1. Species of Arctotidinae investigated. Lewin’s (1922a) infrageneric classification of Arctotis and Beauverd’s
(1915) infrageneric classification of Haplocarpha are followed. The informal species group to which each species is
assigned in this paper is given in parentheses.
Genus Subgenus/section Species
Arctotheca J. C.
Wendl. (4 of 5 spp.
sampled)
A. calendula (L.) Levyns (IV)
A. calendula (L.) Levyns
3
A. populifolia
(P. J. Bergius) Norl. (IV)
A. forbesiana (DC.) K. Lewin (IV)
A. populifolia (P. J. Bergius) Norl. (IV)
A. prostrata (Salisb.) Britten (IV)
Arctotis L. (25 of ca.
60–70 spp. sampled)
sect. Acaules K. Lewin
sect. Acuminatae K. Lewin
sect. Adpressae K. Lewin
sect. Aequales K. Lewin
sect. Angustifoliae K. Lewin
A. acaulis L. (XI)
A. semipapposa (DC.) Beauverd (X)
A. verbascifolia Harv. (XI)
A. campanulata DC. (XIII)
A. angustifolia L. (XIII)
sect. Anomalae K. Lewin
sect. Argentae K. Lewin
sect. Asperae K. Lewin
sect. Austro-orientales K. Lewin
A. dregei Turcz. (V)
A. argentea Thunb. (XIII)
A. aspera L. var. aspera (XIII)
A. arctotoides (L. f.) O. Hoffm. (VII)
A. hispidula (Less.) Beauverd (VII)
sect. Caudatae K. Lewin
sect. Cuneatae K. Lewin
A. microcephala (DC.) Beauverd (VII)
A. perfoliata (L. f.) O. Hoffm. (VIII)
A. suffruticosa K. Lewin (VII)
A. fastuosa Jacq. (VII)
A. cuneata DC. (XIII)
sect. Hirsutae K. Lewin
sect. Leptorhizae K. Lewin
sect. Revolutae K. Lewin
A. rotundifolia K. Lewin (XIII)
A. hirsuta (Harv.) Beauverd (IX)
A. breviscapa Thunb. (VI)
A. cuprea Jacq. (XIII)
A. graminea K. Lewin (XII)
sect. Stoechadinae K. Lewin
Not assigned to a section:
A. venusta Norl. (XIII)
A. gumbletonii Hook. f. (XIII)
A. maidenii Beauverd (VIII)
A. scapiformis Thell. (VII)
A. scullyi Du¨mmer (XIII)
Cymbonotus Cass.
(2 of 2 spp. sampled)
C. lawsonianus Gaudich. (VII)
C. preissianus Steetz (VII)
Haplocarpha Less.
(9 of 9 spp. sampled)
subg. Haplocarpha
subg. Landtia (Less.) Beauverd
H. lanata Less. (II)
H. lyrata Harv. (II)
H. parvifolia (Schltr.) Beauverd (II)
H. scaposa Harv. (II)
H. nervosa (Thunb.) Beauverd (I)
H. rueppellii (Sch. Bip.) Beauverd (I)
H. schimperi (Sch. Bip.) Beauverd (III)
Not assigned to a subgenus: H. hastata K. Lewin (I)
H. oocephala (DC.) Beyers (II)
The uniseriate pappus may comprise 8 free scales
(M. Thulin 1609 [ETH]) or be coronate with all
scales fused in the lower 2/3 (N. P. Barker 1899
[ETH, GRA]). The pappus is up to 1/4 the cypsela
length.
SPECIES GROUP IV
Arctotheca calendula (Fig. 2D–F), A. forbesiana, A.
prostrata (Fig. 2G, H), A. populifolia (Fig.
3A–C).
The species group consists of the four Arctotheca
species examined. The cypselae are obovoid-ellip-
tic, tangentially flattened, with 2 or 3 abaxial ridges
and 2 adaxial ridges, the cypsela surface is trans-
versely rugose with septate trichomes, when present
the coma is composed of twin hairs or multicellular
biseriate trichomes, but neither is very distinct from
the septate trichomes, and the pappus, when pre-
sent, forms a single whorl and is usually shorter
than the cypsela. The cypselae are more strongly

574 Annals of the
Missouri Botanical Garden
Figure 1. Cypselae of Arctotidinae. —A. Haplocarpha hastata, J. W. Ash 1700, abaxial surface. Scale bar
5
500
m
m. —B. H. nervosa, R. A. Dyer 786, abaxial surface. Scale bar
5
500
m
m. —C. H. rueppellii, M. Thulin, A. Hunde
& M. Tadesse 3715, abaxial surface. Scale bar
5
500
m
m. —D. H. lyrata, R. J. McKenzie 1116, abaxial surface (twin
hairs partially removed). Scale bar
5
1 mm. —E. H. lanata, F. J. Kruger 441, abaxial surface (twin hairs partially
removed). Scale bar
5
1 mm. —F. H. lanata, F. J. Kruger 441, adaxial surface. Scale bar
5
1 mm. —G. H. oocephala,
E. E. Esterhuysen 34661, abaxial surface. Scale bar
5
1 mm. —H. Haplocarpha scaposa, R. J. McKenzie 909/4, abaxial
surface. Scale bar
5
1 mm.

Volume 92, Number 4
2005
575McKenzie et al.
Morphology of Cypselae
Figure 2. Cypselae of Arctotidinae. —A. Haplocarpha schimperi, N. P. Barker 1899 (GRA), abaxial surface. Scale
bar
5
500
m
m. —B. H. schimperi, N. P. Barker 1899 (GRA), adaxial surface. Scale bar
5
500
m
m. —C. H. schimperi,
N. P. Barker 1899 (GRA), pappus. Scale bar
5
200
m
m. —D. Arctotheca calendula, R. J. McKenzie 808/3, abaxial
surface. Scale bar
5
500
m
m. —E. A. calendula, R. J. McKenzie 808/3, abaxial surface (trichomes removed). Scale
bar
5
500
m
m. —F. A. calendula, P. A. Bean 2992, abaxial surface. Scale bar
5
500
m
m. —G. A. prostrata, R. J.
McKenzie 1129, abaxial surface. Scale bar
5
500
m
m. —H. A. prostrata, R. J. McKenzie 1129, abaxial surface (trichomes
removed). Scale bar
5
500
m
m.

576 Annals of the
Missouri Botanical Garden
Figure 3. Cypselae of Arctotidinae. —A. Arctotheca populifolia, R. J. McKenzie 771, abaxial surface (trichomes
removed). Scale bar
5
500
m
m. —B. A. populifolia, R. J. McKenzie 771, adaxial surface (trichomes removed). Scale
bar
5
500
m
m. —C. A. populifolia, R. J. McKenzie 771, pappus. Scale bar
5
500
m
m. —D. Arctotis dregei, R. J.
McKenzie 834/1, abaxial surface. Scale bar
5
1 mm. —E. A. dregei, R. J. McKenzie 815/2, abaxial surface. Scale bar
5
1 mm. —F. A. dregei, R. J. McKenzie 815/2, adaxial surface. Scale bar
5
1 mm. —G. Arctotis breviscapa, C. Boucher
& P. Shepherd 460, abaxial surface. Scale bar
5
1 mm. —H. A. breviscapa, J. A. Marsh 1214, adaxial surface. Scale
bar
5
500
m
m.

Volume 92, Number 4
2005
577McKenzie et al.
Morphology of Cypselae
convex on the abaxial surface than the adaxial sur-
face, particularly in A. populifolia. The position of
the prominent ridges is variable; the lateral abaxial
ridges may be inconspicuous in A. calendula; the
median abaxial ridge may be inconspicuous in A.
calendula (Fig. 2F), A. populifolia, and A. prostrata.
Septate trichomes often form a dense woolly indu-
mentum completely concealing the cypsela and
pappus if present, but in some A. calendula and A.
prostrata specimens the indumentum is sparse. A
coma of multicellular biseriate trichomes is present
in A. calendula. In A. forbesiana and A. prostrata
the coma comprises long, thread-like twin hairs.
The coma is about 1/5 of the cypsela length in A.
populifolia and is of similar length to the cypsela
in the other species. A pappus may be absent in
A. calendula (Fig. 2F) and A. prostrata. In A. ca-
lendula the pappus, when present, is of two forms:
either 2 lateral obovate to oblanceolate scales with
often a third much reduced adaxial scale; or a
whorl of 7 to 9 scales of similar length (e.g., C.
Cottrell 7 [GRA]). The pappus comprises a whorl
of 8 or 9 scales in A. forbesiana and a whorl of 7
to 9 scales when present in A. prostrata. The pap-
pus is always coronate in A. populifolia with 7 to
11 lobes, but the degree of scale fusion is variable;
some scales may be completely free but are usually
fused in at least the basal half. The pappus is very
reduced in A. populifolia (0.25–0.5 mm long, up to
0.15
3
cypsela length). In the other species the
pappus consists of scales to 1.3 mm long; in A.
prostrata the scales may be of similar length to the
cypsela, but are usually more reduced. In all spe-
cies the rim of the apical plate forms a shallow,
lobed cup around the base of the pappus (Fig. 2E,
3C).
SPECIES GROUP V
Arctotis dregei (Fig. 3D–F)
This species is assigned to Arctotis sect. Anom-
alae. The cypselae are obovoid-elliptic, tangentially
flattened, with 2 wide-spreading abaxial wings cre-
ating a single elliptic-sinuate to linear-sinuate ab-
axial cavity, 2 short adaxial wings (most evident in
transverse section), a coma of twin hairs, and a pap-
pus of scales in 2 whorls. In some specimens a
weakly raised median abaxial ridge is present in
the cavity, but is never developed into a wing. In
some specimens the abaxial cavity is narrow (Fig.
3D), but usually the cavity is very broad (Fig. 3E).
The abaxial wings are transversely rugose and bear
rounded-truncate teeth; the adaxial wings are not
rugose or dentate. Unicellular trichomes are dense
on the abaxial wings and sparse to dense within the
cavity. On the adaxial surface uniseriate trichomes
are usually moderate-dense on the upper part and
the adaxial wings, and glabrous in the center and
lower part; in some specimens the adaxial surface
is entirely glabrous. The coma comprises a ring of
twin hairs, 1/5 to equal the cypsela length, notably
longer on the adaxial surface. The inner pappus
whorl of 8 or 9 scales is longer than the outer whorl
and varies in length from 0.5 to 1.5 times the cyp-
sela length. The outer whorl consists of 6 to 9
scales; when the whorl is incomplete, scales are
absent on the abaxial side. The scale apex is obtuse
to rounded in the inner whorl and acute to obtuse
in the outer whorl.
SPECIES GROUP VI
Arctotis breviscapa (Fig. 3G, H)
This species is assigned to Arctotis sect. Leptor-
hizae. The cypselae are unique in having 2 well-
developed adaxial wings, in addition to 3 abaxial
wings. The cypselae are oblanceolate-compressed,
with a short coma of twin hairs and a uniseriate
pappus of 8 scales. The lateral abaxial wings are
strongly incurved, usually occluding the abaxial
cavities and concealing the median abaxial wing.
Both the abaxial and adaxial wings are entire. The
adaxial cavities are linear. Dense, short septate tri-
chomes to 100
m
m long are present on the abaxial
and lateral surfaces. Dense unicellular trichomes
cover the adaxial surface between the adaxial
wings. Papillose trichomes are present on both sur-
faces in some specimens (e.g., C. Boucher & P.
Shepherd 4602 [NBG, PRE]). The coma is up to 1/4
the cypsela length. The pappus is at most equal in
length to the cypsela. The pappus scale apex is
rounded.
SPECIES GROUP VII
Arctotis arctotoides (Fig. 4A, B), A. fastuosa (Fig.
4C, D), A. hispidula, A. microcephala (Fig. 4E,
F), A. scapiformis (Fig. 4G, H), A. suffruticosa,
Cymbonotus lawsonianus (Fig. 5A, B), C. preis-
sianus (Fig. 5C, D)
This group contains five species assignable to
Arctotis sect. Austro-orientales, A. fastuosa, and the
two species of Cymbonotus. The cypselae in this
group have 2 or 3 abaxial wings, a single linear or
sinuate-linear abaxial cavity, simple uniseriate tri-
chomes or biseriate glandular trichomes, a coma of
twin hairs is absent, and a pappus is present in two
forms of A. fastuosa. Two adaxial ridges are usually
present; one or additional adaxial ridges may occur
in A. microcephala and A. scapiformis. The cypselae

578 Annals of the
Missouri Botanical Garden
Figure 4. Cypselae of Arctotidinae. —A. Arctotis arctotoides, R. J. McKenzie 724/1, abaxial surface. Scale bar
5
500
m
m. —B. A. arctotoides, R. J. McKenzie 724/1, adaxial surface. Scale bar
5
500
m
m. —C. A. fastuosa, J. Hutchinson
889, abaxial surface. Scale bar
5
200
m
m. —D. A. fastuosa, J. Hutchinson 889, adaxial surface. Scale bar
5
200
m
m.
—E. A. microcephala, W. J. Hanekom 1963, abaxial surface. Scale bar
5
500
m
m. —F. A. microcephala, W. J. Hanekom
1963, adaxial surface. Scale bar
5
500
m
m. —G. A. scapiformis, A. White 41, abaxial surface. Scale bar
5
500
m
m.
—H. A. scapiformis, A. White 41, adaxial surface. Scale bar
5
500
m
m.

Volume 92, Number 4
2005
579McKenzie et al.
Morphology of Cypselae
Figure 5. Cypselae of Arctotidinae. —A. Cymbonotus lawsonianus, J. M. Dalby 86/128, abaxial surface. Scale bar
5
500
m
m. —B. C. lawsonianus, J. M. Dalby 86/128, adaxial surface. Scale bar
5
500
m
m. —C. C. preissianus, H.
Salasoo 2113, abaxial surface. Scale bar
5
500
m
m. —D. C. preissianus, H. Salasoo 2113, adaxial surface. Scale bar
5
500
m
m. —E. Arctotis maidenii, T. J. Whaite 3787, abaxial surface. Scale bar
5
1 mm. —F. A. maidenii, T. J.
Whaite 3787, adaxial surface. Scale bar
5
1 mm. —G. A. perfoliata, W. E. Dix 156, abaxial surface. Scale bar
5
500
m
m. —H. A. perfoliata, W. E. Dix 156, adaxial surface. Scale bar
5
500
m
m.

580 Annals of the
Missouri Botanical Garden
Table 2. Selected cypselar morphological characters for Arctotidinae examined. See Appendix 1 for key to the
characters. The informal species group to which each species is allocated in this paper is given in parentheses. NA
5
character is not applicable.
Species 123456789101112
Arctotheca calendula (IV)
A. forbesiana (IV)
A. populifolia (IV)
A. prostrata (IV)
Arctotis acaulis (XI)
3
2
3
3
5
1
1
1
1
0
0
0
0
0
1
3
3
3
3
0
1
1
0
0
1
1
1
1
1
0/1
1
2
1
1
2/3
0
0
0
0
1
0
0
0
0
2
0
0
0
0
0
NA
NA
NA
NA
7
NA
NA
NA
NA
1/2
A. angustifolia (XIII)
A. arctotoides (VII)
A. argentea (XIII)
A. aspera var. aspera (XIII)
A. breviscapa (VI)
1
2
1
1
4
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
1
2
1
1
1
0
0/1
0
0
0
2/3
1
2
2
1
1
1
1
1
2
2
1
2
2
0
0
0
0
0
2
6
2
4
2
NA
1
1/2
1/2
2
1
A. campanulata (XIII)
A. cuneata (XIII)
A. cuprea (XIII)
A. dregei (V)
A. fastuosa (VII)
A. graminea (XII)
1
1
1
3
2
4
0
1
0
1
1
0
0
0
0
0
0
0
0
0
0
1
3
0
1
1
1
0/1
0
1
0
0/1
0
0
0
0
2
2
2
1
1
1
1
1
1
2
1
1
2
2
2
1
1
2
0
0
0
2
0
0
3
3/4
3
1/2
2
3
1
1/2
2
2
2
1
A. gumbletonii (XIII)
A. hirsuta (IX)
A. hispidula (VII)
A. maidenii (VIII)
A. microcephala (VII)
1
2
2
2
2
0
0
0
1
0
0
1
0
0
0
0
1
3
1/2
3
1
1
2
1
2
0
1
0/1
0
1
2
3
2
1
1
1
1
1
1
1
2
2
2
1
1
0
0
0
0
0
3
6
3/4
5
2
2
1/2
1/2
1/2
1
A. perfoliata (VIII)
A. rotundifolia (XIII)
A. scapiformis (VII)
A. scullyi (XIII)
A. semipapposa (X)
4
1
2
1
5
0
0
0
0
0
0
0
0
0
0
3
0
2
0
0
1
1
2
1
1
0/1
1
1
0
0
1/2
2
1
1/2
2
1
1
1
1
1
1/2
2
1
2
2
0
0
0
0
0
2/5
3
2
2/4
2
1
1/2
1
1/2
2
A. suffruticosa (VII)
A. venusta (XIII)
A. verbascifolia (IX)
C. lawsonianus (VII)
2
2
4
2
0
0
0
0
0
0
1
0
3
0/1
0
3
2
1
1
2
1
0
0
0
1
2
2
1
1
1
1
1
1
2
2
1
0
0
0
0
2
2
3
2
1
1/2
1
2
C. preissianus (VII)
H. hastata (I)
H. lanata (II)
H. lyrata (II)
H. nervosa (I)
2
2
5
5
2
0/1
1
0
0
0
0
1
0
0
0
3
0
0
0
0
2
1
1
1
1
1
0
0
0
0
1
2
2
2
2
1
0
0
1
0
1
0
0
0
0
0
0
0
0
0
2
NA
NA
2
NA
2
NA
NA
1
NA
H. oocephala (II)
H. parvifolia (II)
H. rueppellii (I)
H. scaposa (II)
H. schimperi (III)
1
1
2
1
2
0
0
0
0
0
0
0
1
0
0
1
0
0
0
3
2
2
1
1
2
1
1
0
0
1
2
2
2
2
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
are obconical in A. fastuosa and obovoid or ob-
ovoid-cylindrical in the other species. The cypsela
is radially curved with a concave abaxial surface
and convex adaxial surface in some specimens of
A. fastuosa, A. hispidula, and C. lawsonianus. The
surface is transversely rugose only on the abaxial
surface in A. scapiformis and some specimens of A.
fastuosa, and on both abaxial and adaxial surfaces
in the other species. The abaxial wings are only
weakly to moderately developed and strongly in-
curved, and the median wing or ridge is always
shorter than the lateral ones. The abaxial wings are
least developed in a form of A. fastuosa (Fig. 4C).
The wings may be entire or dentate and the shape
of the teeth ranges from acute to truncate (Table 2).
Teeth were not observed in A. microcephala and A.
scapiformis. Only simple unicellular or uniseriate
trichomes (to 30
m
m long) are present on the cyp-
sela surface; papillose trichomes are common in C.
preissianus. A coma of twin hairs is absent in all
specimens examined. Instead, one or a few biser-
iate glandular trichomes are present on the carpo-

Volume 92, Number 4
2005
581McKenzie et al.
Morphology of Cypselae
Table 2. Extended.
13 14 15 16 17 18 19 20 21 22
NA
NA
NA
NA
1
NA
NA
NA
NA
1
1
1
1
1
1
2/3
3
2/3
2/3
0/2/3
3
3
2/3
2/3
0/2/3
1
1
1
1
1
1
1–1.2
0.2
1
1–1.4
1/2
1
1
1
1
0
0
0
0
1
0/1
1
1
0/1
2
NA
3/4
1
2/3
NA
1
2
2
2
3
1/5
2
1
1
2/5
1/2
0/1/2
3
3
3
3
0/1/2
3
3
3
1
2
1
0
1
1.25–1.4
NA
1–1.5
1.2–1.5
0.1–0.25
1
3
1
1
1
1
0
1
0
0
2
0
2
2
1
NA
1
1
2/3/4
2/3
NA
1
2
1
1/2
1
1
1
1
1
2
2
1
1/2
0/2
3
3
1/2
2
3
3
3
0/1
2/3
1
1
1
1
1
0
1
0.8–1.1
0.75–1.3
1.6–1.9
0.75–1
NA
0.15–0.3
1
1
1
1
NA
1
1
1
1
1
1
1
2
2
2
2
0
2
1/2
2
1
1/2
NA
1
1
2
3
2
NA
6
2
2
2
NA
1
1/2
0
1
NA
3
1/2
1/2
1
1
0
2
2
2
1–1.5
NA
NA
NA
NA
1
NA
3
3
3
0
1
0
0
0
2
1
0
0
0
NA
1/2/3
NA
2/3
1/2/3
2
2
2
2/3
1
NA
1
NA
1
1
NA
0/3
NA
1/3
1
NA
3
NA
3
1
0
1
2
1
1/2
NA
0.3–0.5
NA
1.5–2
,
0.05
NA
1
3
1
3
0
1
0
1
0
0
2
0
2
1/2
NA
1
NA
2/3
2
1
1
2
2
1
1
2
2
1/2
2/3
1
2
3
3
3
2
1
1
0
NA
1.25–1.5
1.25–1.3
NA
3
1
1
NA
0
1
1
0
0
2
2
0
1/2/3
NA
NA
NA
NA
2
NA
NA
1
NA
2
NA
3
3
2/4
3
0
2
3
2/3
3
0
2
3
2/3
0/3
0
1
1
1
NA
NA
0.75–1
0.75–1
0.1–0.15
3
NA
1
1
1
0
0
0
0
0
0
1
1
1
2
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
4
4
NA
NA
2
1/2
2
0
0
1
0
0
0
0
1
1
1
0
1
0
1.5–1.75
1.75–2
NA
3.5
NA
1
1
NA
1
NA
1
1
0
0
0
2
2
2
2
1
podium on the abaxial surface (usually between the
abaxial wings) in A. sect. Austro-orientales species
or on the adaxial surface in some C. preissianus
specimens. No trichomes were observed on the car-
popodium in A. fastuosa and C. lawsonianus. In A.
fastuosa the pappus, when present, consists of ei-
ther a single whorl of 2 to 7 scales to 0.15 mm
long, or 2 whorls of 1 to 4 and 7 to 9 scales re-
spectively, with the longest scales (to 0.3 mm long)
on the adaxial surface.
SPECIES GROUP VIII
Arctotis maidenii (Fig. 5E, F), A. perfoliata (Fig.
5G, H)
Based on cypselar morphology Arctotis perfoliata
is anomalous within A. sect. Austro-orientales; A.
maidenii has not been assigned to a section. Both
species have large obovoid, tangentially flattened
cypselae with 3 abaxial wings and 2 adaxial ridges,
and lack a coma of twin hairs and a pappus. The

582 Annals of the
Missouri Botanical Garden
Table 2. Extended.
23 24 25 26 27 28 29
0
0
0
0
8
2–9
8–9
7–11
7–9
8
NA
NA
NA
NA
2
0.2–0.7
0.25–0.3
0.05–0.15
0.4–1
1–1.9
0
0
1
0
0
1
1
NA
1
2
1/2
2
3
1/2
3
c.5
NA
8
8
0
8
NA
8
8
8
2
NA
2
2
NA
0.35–0.85
NA
1–1.5
1.5–1.7
0.75–1
0
NA
0
0
0
1
NA
2
2
2
2/3
NA
3
2/3
3
8
8
8
6–9
NA
8
8
8
8
8–9
NA
8
2
2
2
2
NA
2
1.25–1.4
1.25–1.9
1.7–1.9
0.5–1.6
NA
0.3
0
0
0
0
NA
0
2
2
2
1
NA
2
2/3
2/3
2/3
2/3
NA
3
8
0
NA
NA
NA
8
8
NA
NA
NA
2
NA
NA
NA
NA
1.5–2
0.04–0.06
NA
NA
NA
0
0
NA
NA
NA
2
1
NA
NA
NA
3
3
NA
NA
NA
NA
8
NA
8
0–c.10
NA
8
NA
8
5–8
NA
2
NA
2
2
NA
1.25–1.5
NA
1.25–1.7
0.03–0.25
NA
0
NA
0
0
NA
2
NA
2
1
NA
3
NA
2/3
2/3
NA
8
8
NA
NA
8
8
NA
NA
2
2
NA
NA
1–1.5
1.25–1.8
NA
NA
0
0
NA
NA
2
2
NA
NA
2
2/3
NA
NA
NA
0
0
8
NA
NA
9
8
8
NA
NA
NA
NA
1
NA
0.1
1–1.4
1.5–1.9
,
0.1
NA
1
0
0
0
NA
NA
1
1
2
NA
NA
1
0
2/3
8
8
ca. 8
6–7
0
8
8
ca.8
8
7–9
2
2
1
2
NA
1.3–1.7
1.3–1.5
0.05
4.2–4.6
0.15–0.25
0
0
0
0
0/1
2
2
2
1
1/2
2/3
2/3
1
0
1/2
abaxial cavities are often linear or linear-sinuate,
but in some specimens are almost occluded by the
strongly incurved wings (Fig. 5E). Both cypsela sur-
faces may be transversely rugose in both species,
but more strongly in A. perfoliata. In A. maidenii
sparse or moderately dense simple uniseriate tri-
chomes to 60
m
m long are present on the median
abaxial wing and adaxial surface, and a tuft of glan-
dular biseriate trichomes is present on the carpo-
podium between the abaxial wings. The cypselae
are glabrous in A. perfoliata. The cypselae of A.
maidenii are the longest (5–6 mm long) of any spe-
cies examined.
SPECIES GROUP IX
Arctotis hirsuta (Fig. 6A, B)
Arctotis hirsuta is assigned to A. sect. Hirsutae.
The cypselae are small (to 2.75 mm long) with 2
ovate-linear abaxial cavities, simple uniseriate tri-
chomes, a coma is absent, and minute pappus
scales are present. The apical plate is offset to the

Volume 92, Number 4
2005
583McKenzie et al.
Morphology of Cypselae
Figure 6. Cypselae of Arctotidinae. —A. Arctotis hirsuta, N. S. Pillans 7693, abaxial surface. Scale bar
5
500
m
m. —B. A. hirsuta, N. S. Pillans 7693, adaxial surface. Scale bar
5
500
m
m. —C. Arctotis semipapposa, R. J.
McKenzie 890, abaxial surface. Scale bar
5
500
m
m. —D. A. semipapposa, R. J. McKenzie 890, adaxial surface. Scale
bar
5
500
m
m. —E. A. acaulis, R. J. McKenzie 848/1, abaxial surface (trichomes partially removed). Scale bar
5
1
mm. —F. A. graminea, R. J. McKenzie 1084/2, abaxial surface. Scale bar
5
1 mm. —G. A. graminea, R. J. McKenzie
1084/2, adaxial surface. Scale bar
5
1 mm. —H. A. angustifolia, E. E. Esterhuysen 31373, abaxial surface (coma
partially removed). Scale bar
5
1 mm.

584 Annals of the
Missouri Botanical Garden
Figure 7. Cypselae of Arctotidinae. —A. A. aspera var. aspera, R. J. McKenzie 847/1, abaxial surface (trichomes
removed). Scale bar
5
1 mm. —B. A. cuneata, R. J. McKenzie 826/1, abaxial surface. Scale bar
5
1 mm. —C. A.
cuneata, R. J. McKenzie 826/1, adaxial surface. Scale bar
5
1 mm. —D. A. cuprea, R. Schlechter 8697, abaxial surface
(twin hairs and trichomes partially removed). Scale bar
5
1 mm. —E. A. gumbletonii, J. W. Mathews s.n., abaxial
surface. Scale bar
5
1 mm. —F. A. rotundifolia, E. E. Esterhuysen 11129, abaxial surface. Scale bar
5
1 mm. —G.
A. scullyi, H. H. Bolus 9615, abaxial surface. Scale bar
5
1 mm. —H. A. venusta, R. J. McKenzie 878/1, abaxial
surface. Scale bar
5
1 mm.

Volume 92, Number 4
2005
585McKenzie et al.
Morphology of Cypselae
abaxial side, and the cypsela has a strongly convex
adaxial surface. The median abaxial wing is longer
than the lateral abaxial wings. The cavities are
rounded to obtuse at the apex and base. The 3 ab-
axial wings are fused in the lower part above the
carpopodium. The base of the cypsela is short and
constricted. The abaxial wings are entire or bear
short, narrow, obtuse teeth; the wings are barely to
strongly transversely rugose. The adaxial surface is
always smooth. Clavate trichomes are sparse to
moderately dense on the abaxial surface and dense
on the adaxial surface. The cypselae have a unis-
eriate pappus of 8 or 9 rounded scales to 0.15 mm
long, which are often longer on the adaxial surface.
SPECIES GROUP X
Arctotis semipapposa (Fig. 6C, D)
The species is placed in Arctotis sect. Acumi-
natae. The cypselae are obovate-cylindrical with a
broad base, 3 abaxial wings forming 2 linear-sin-
uate cavities, and 2 adaxial ridges. The cypsela sur-
face is smooth and bears sparse septate trichomes
100–135
m
m long with an acute apex. The wings
have short acute to rounded teeth. A coma of twin
hairs is absent. Instead, glandular biseriate tri-
chomes are usually present on the carpopodium,
forming either a complete ring or tufts between the
abaxial wings and adaxial ridges. The pappus is
usually a single, incomplete whorl of 5 to 8 scales
to 0.5 mm long. The scales are hemispherical with
a rounded apex. The scales are longer on the ab-
axial side; adaxial scales are either rudimentary or
absent. A partial outer whorl of rudimentary scales
may be present on the abaxial side.
SPECIES GROUP XI
Arctotis acaulis (Fig. 6E), A. verbascifolia
The two species are placed in Arctotis sect. Acau-
les and Adpressae respectively. The cypselae are ob-
conical-obovoid, with 3 abaxial wings and 2 adaxial
ridges, the wings are fused in the lower part, the
cypsela base is constricted into a short ‘‘stalk,’’ sep-
tate trichomes and a coma of twin hairs are present,
and the pappus of free scales is biseriate. The cyp-
selar surface is smooth. The wings are entire or
bear narrow, acute teeth on the lower half. The ab-
axial cavity is obovate or obovate-sinuate in A.
acaulis and oblanceolate in A. verbascifolia. Septate
trichomes are usually moderate to dense on the ab-
axial and adaxial surfaces. The coma exceeds the
cypsela in length. Each pappus whorl consists of 8
scales. The outer whorl is much shorter than the
inner whorl. The scales of the inner whorl are lon-
gest on the abaxial side, whereas the scales of the
outer whorl are longest on the adaxial side. The
scale apex is rounded in the inner whorl and obtuse
in the outer whorl. The pappus exceeds the cypsela
in length.
SPECIES GROUP XII
Arctotis graminea (Fig. 6F, G)
The species is assigned to Arctotis sect. Revolu-
tae. The obconical-obovoid cypselae are distin-
guished by the reduced biseriate pappus, the short
coma, and covering of short septate trichomes to
about 300
m
m long on all surfaces. Three abaxial
wings and two adaxial ridges are present. The me-
dian abaxial wing is shorter than the lateral wings,
which are entire and shortly inflexed. The abaxial
cavity is linear-cuneiform. The cypselar surface is
smooth. A short coma of twin hairs to 1 mm long
(1/3 of the cypsela length) is present. Each pappus
whorl consists of 8 scales with a rounded apex. The
inner pappus whorl is up to 1.8 mm long (1/3 of
the cypsela length); the scales of the outer whorl
are rudimentary.
SPECIES GROUP XIII
Arctotis angustifolia (Fig. 6H), A. argentea, A. as-
pera var. aspera (Fig. 7A), A. campanulata, A.
cuneata (Fig. 7B, C), A. cuprea (Fig. 7D), A.
gumbletonii (Fig. 7E), A. rotundifolia (Fig. 7F),
A. scullyi (Fig. 7G), A. venusta (Fig. 7H)
This species group comprises the remaining 11
Arctotis species examined. The species are assigned
to seven sections: Angustifoliae, Argentae, Asperae,
Caudatae, Cuneatae, Revolutae, and Stoechadinae.
The cypselae are variably obconical with 3 well-
developed abaxial wings and 2 adaxial ridges, sep-
tate trichomes are present on the cypsela surface,
a coma of twin hairs is always present, and the
pappus of scales is biseriate. The cypsela is obcon-
ical, obconical-obovoid or obconical-cylindrical in
different species. The cypselar surface is smooth.
The wings are entire or dentate; this may vary with-
in a species. In some species the wings almost oc-
clude the abaxial cavities (Fig. 7G). The median
abaxial wing is shorter than, similar in length to,
or longer than the lateral wings in different species.
The abaxial cavity is either linear (A. venusta), lin-
ear-obovate (e.g., A. angustifolia), triangular-pan-
durate or linear-cuneiform. Apart from the coma,
the cypselae may be entirely glabrous, as in A.
gumbletonii (Fig. 7E). Usually septate trichomes
are present on the cypsela surface; the density
ranges from sparse to dense on both abaxial and

586 Annals of the
Missouri Botanical Garden
adaxial surfaces and may vary within a species (Ta-
ble 2) and between the ray and disc florets. The
terminal cell may be long, thread-like and inter-
woven with other trichomes to produce a woolly in-
dumentum (Fig. 7C, G) or very reduced (less than
150
m
m long) and not interwoven (e.g., A. angus-
tifolia). In some species, such as A. aspera and A.
laevis, the septate trichomes conceal the abaxial
cavities. The coma consists of twin hairs varying in
length from up to 1/3 the cypsela length to about
1.5 times the cypsela length. The coma is decidu-
ous in A. angustifolia and persistent in the other
species. The outer pappus whorl is much shorter
than the inner whorl. In A. angustifolia the abaxial
scales of the outer whorl may be reduced to tufts
of minute bristles. The scales of the inner whorl are
longest on the abaxial side, whereas the scales of
the outer whorl are longest on the adaxial side. The
scale apex is rounded in the inner whorl and obtuse
in the outer whorl. The pappus length varies from
slightly shorter than the cypsela length in A. an-
gustifolia (Fig. 6H) to twice the cypsela length in
A. gumbletonii (Fig. 7E).
D
ISCUSSION
In examining the SEM micrographs, variable and
notable features of the cypselae are apparent: lon-
gitudinal ridges or wings on the abaxial surface, the
pappus scales, and surface features.
ABAXIAL RIDGES OR WINGS
Cypselae of Arctotidinae present a progression
from species with only weakly developed abaxial
ridges (our Group I), through species with conspic-
uous abaxial ridges (Groups II–IV), to species with
well-developed abaxial wings (Groups V–XIII). The
number and extent of development of the abaxial
ridges is constant within each species group with
only few exceptions. The number of abaxial ridges
is variable in Arctotheca calendula and A. prostrata
(IV), and in Arctotis microcephala and A. scapifor-
mis (VII) two to four adaxial ridges may be con-
spicuous. In Haplocarpha oocephala and H. parvi-
folia (II), the abaxial ribs are broad and separated
by narrow furrows; in other species of Group II the
ribs are narrow.
In Arctotis and Cymbonotus (V–XIII), two or
three abaxial wings are developed that form either
one or two cavities. The presence of the abaxial
wings is a diagnostic feature of these genera; wings
are more strongly developed in Arctotis (Bremer,
1994). Variation in development of the abaxial
wings, shape of the cavities, and surface features
(e.g., the presence of teeth and rugose ridges) af-
fords means of distinguishing informal species
groups. The median abaxial wing may be poorly
developed so that only one cavity is obvious (e.g.,
A. dregei [V] and A. fastuosa [VII]). The lateral ab-
axial wings are less developed in Group VII than
in Groups VIII–XIII. In A. breviscapa (VI) the two
adaxial ridges are extended to form distinct wings;
adaxial wings are also present, but less developed,
in A. dregei. Lewin (1922a) emphasized this char-
acteristic to distinguish A. sect. Leptorhizae; clearly,
the presence of adaxial wings per se is not unique
to A. breviscapa. Degree of inflexion of the lateral
abaxial wings and development of the median ab-
axial wing are very different in A. breviscapa and
A. dregei and are distinctive within Arctotis. In
some species dentation of the abaxial wings is var-
iable, particularly in length and shape, and is less
helpful for delimiting species groups.
Outside of Arctotis and Cymbonotus, H. lyrata (II)
is the only species in which the lateral abaxial ridg-
es are curved toward the median abaxial ridge. This
feature was noted also by Reese (1989). The lateral
ridges are not well enough developed to be consid-
ered wings and do not create distinct collateral cav-
ities as in Arctotis and Cymbonotus. This may rep-
resent either an early stage in the evolution of
wings, or perhaps a past hybridization event with
an Arctotis species. In other external cypsela fea-
tures, H. lyrata is closely allied to H. lanata and,
anatomically, the cypselae of H. lyrata are similar
to Arctotheca calendula, Arctotis, and C. lawsoni-
anus (Reese, 1989).
PAPPUS
The pappus is variable in Arctotidinae and is
helpful to distinguish some species groups. When
a pappus is present, it usually comprises unfused
scales. A coronate pappus was observed in three
species only (Arctotheca populifolia [IV], Haplocar-
pha hastata [I], and H. schimperi [III]), and in A.
populifolia and H. schimperi the degree of fusion of
the scales varies. A pappus of free scales is best
developed in Groups II and XIII. The absence or
extreme reduction of the pappus was used to sup-
port the segregation of Landtia from Haplocarpha
(Phillips, 1951) and of Venidium from Arctotis (Les-
sing, 1832). The pappus is easily lost or reduced
in some Compositae, so its form is not always a
reliable generic criterion. The pappus is particu-
larly variable in Arctotheca (IV—specimens may
have no pappus, a single whorl of two to nine free
scales, or a variably fused coronate pappus). The
absence of even vestigial pappus scales is charac-
teristic of most species in Groups VII and VIII, as

Volume 92, Number 4
2005
587McKenzie et al.
Morphology of Cypselae
well as some specimens of A. calendula and A. pros-
trata (IV). Minute scales are characteristic of Group
I, Arctotis hirsuta (IX), and some forms of A. fas-
tuosa (VII). Previously, Arctotis semipapposa (X)
was thought to be characterized by a very reduced
uniseriate pappus comprising scales only on the ab-
axial side of the cypsela (Harvey, 1865); specimens
examined in the present study indicate the number
of scales is variable and that minute adaxial scales
and a partial second whorl may also be present.
The scale apex shape is often variable within a
species group and between the whorls on a single
cypsela; acuminate apices are found only in H. sca-
posa and H. lyrata (II).
CYPSELA TRICHOMES
Cypselar pubescence was emphasized to support
segregation of Cryptostemma and Microstephium
from Arctotheca (e.g., Harvey, 1865; Bentham,
1873a, b); it was subsequently found to be unten-
able at the generic level (Lewin, 1922a). Pubes-
cence density may vary between populations in Arc-
totheca and Arctotis species, even between ray and
disc cypselae in the same capitulum in some Arc-
totis species; thus, it must be utilized cautiously as
a taxonomic character.
Herman (2001) reported the presence of septate
trichomes in four Arctotheca and three Arctotis spe-
cies, and their absence in three Haplocarpha spe-
cies. Our observations on a wider sample of species
confirm their presence in Arctotheca (IV) and the
‘‘core’’ Arctotis group (XIII), and absence in Cym-
bonotus (VII) and Haplocarpha (I–III). The length
of septate trichomes varies considerably among Arc-
totis species. Short septate trichomes (to 300
m
m
long) characterize the cypselae of A. breviscapa
(VI), A. graminea (XII), and A. semipapposa (X);
septate trichomes are typically much longer,
thread-like, and interwoven to form a woolly indu-
mentum in the ‘‘core’’ Arctotis group (XIII).
The presence of multicellular biseriate trichomes
on cypselae of Arctotheca calendula (IV) was noted
by Herman (2001). This trichome type was not ob-
served in any other species. Herman stated these
trichomes were interspersed among the septate tri-
chomes; in specimens examined in the present
study, multicellular biseriate trichomes were at-
tached at the base of the cypsela only and thus may
represent a coma. As indicated in Figure 2D, the
multicellular biseriate trichomes are similar to the
septate trichomes and require microscopic exami-
nation to distinguish. In all other Arctotidinae, the
coma, when present, is composed of twin hairs.
Biseriate glandular trichomes are present on the
carpopodium in Group VII and A. semipapposa (X),
but rarely form a complete ring. These species lack
a coma of twin hairs.
CORRELATION OF CYPSELAR MORPHOLOGY
WITH TAXONOMY
Arctotheca
The four Arctotheca species studied form a close-
ly knit and easily recognizable group (our Group
IV) based on cypselar morphology. The cypselae
are characterized by being bilaterally flattened with
lateral ridges always well developed, wings absent,
a coma of either twin hairs (A. forbesiana, A. po-
pulifolia, and A. prostrata) or multicellular biseriate
trichomes (A. calendula), a moderate or dense cov-
ering of septate trichomes, and a pappus shorter
than the cypsela (in some specimens the pappus is
wanting). The sole species not studied, A. margin-
ata, has narrowly elliptic, densely pilose cypselae
bearing seven to ten narrowly elliptic, hyaline pap-
pus scales with an acute apex (Beyers, 2002). In
combination with its perennial, stoloniferous habit,
this indicates A. marginata has a close affinity with
A. prostrata. Arctotheca is well supported by other
morphological characteristics, such as the posses-
sion of neuter ray florets, bisexual disc florets, and
papillose filaments.
Cypselar pubescence and pappus structure have
been used to discriminate Arctotheca, Cryptostem-
ma, and Microstephium (e.g., Harvey, 1865; Ben-
tham, 1873a, b; Beauverd, 1915). The reduced im-
portance of these features as generic criteria in this
group was recognized by Lewin (1922a). The pre-
sent study highlights the variability of these fea-
tures in A. calendula and A. prostrata and thus sup-
ports Lewin’s (1922a) amalgamation of the three
genera.
At least three distinct cypsela forms exist in A.
calendula and A. prostrata. The cypselae of these
two species vary in terms of the number of pappus
scales present, number of prominent abaxial ridges,
and the density of septate trichomes. According to
previous authors, A. prostrata lacks a pappus (e.g.,
Candolle, 1837; Harvey, 1865; Lewin, 1922a;
Goldblatt & Manning, 2000; Beyers 2002), or at
most two pappus scales are present (Beauverd,
1915). In the present study, a pappus was absent
in one specimen of A. prostrata examined (P. A .
Bean 2992 [BOL]) from the Touwsberg, Western
Cape province. In other A. prostrata specimens ex-
amined, a uniseriate pappus of 7 to 9 scales was
present. A single specimen of A. calendula exam-
ined (J. Samuel 19 [GRA]) from the western North-
ern Cape province also lacked a pappus. This in-

588 Annals of the
Missouri Botanical Garden
dicates absence of a pappus is not a distinguishing
characteristic of A. prostrata. The variation in cyp-
selar morphology in A. calendula and A. prostrata
indicates a reappraisal of their taxonomy is war-
ranted. Both species also exhibit considerable var-
iation in vegetative morphology (Harvey, 1865).
Arctotheca calendula and A. prostrata can be dis-
criminated on life form: A. calendula is a tufted
annual, A. prostrata is perennial with prostrate, ad-
ventitiously rooting stems. Distinguishing them on
reproductive morphology is less straightforward.
Some forms of A. calendula from the Northern Cape
province possess cypselae morphologically very
similar to those of A. prostrata from the Eastern
Cape province. A closer examination of geograph-
ical variation patterns, particularly of cypselar and
involucral-bract morphology, is needed to clarify
their circumscription.
Arctotheca populifolia cypselae are distinctive in
the genus in their larger dimensions and coronate
pappus. Intraspecific variation is mainly restricted
to the number and extent of fusion of the pappus
scales, as noted by Lewin (1922a). The pappus
scales may be nearly completely separate or almost
completely fused in cypselae from the same capit-
ulum (Lewin 1922a). An additional potential source
of variation is cryptic hybridization. A putative hy-
brid between A. calendula and A. populifolia from
Maccassar Beach, Western Cape province (C. Peter
84 [GRA]) has cypselae of intermediate dimensions
with nine completely free pappus scales, interme-
diate pubescence, and an indistinct coma.
The cypselae of A. forbesiana are very similar to
forms found in A. calendula and A. prostrata. Va r -
iability in this species is unknown, as only one
specimen with mature cypselae was available for
study.
Arctotis dregei (V) strongly resembles Arctotheca
calendula in vegetative and floral morphology, and
the two species are often confused in herbaria. The
ray florets are neuter in both species and in corolla,
filament, and style morphology A. dregei is barely
distinguishable from Arctotheca. With respect to
cypselar morphology, the affinities of A. dregei lie
with Arctotis. A well-developed biseriate pappus of
free scales, well-developed abaxial wings creating
a distinct abaxial cavity, and a distinct coma of twin
hairs are features shared among A. dregei and other
Arctotis species.
Arctotis
On the basis of cypselar morphology Arctotis spe-
cies assort to nine informal species groups herein.
These generally correspond with 5 of Lewin’s
(1922a) 15 infrageneric sections, namely, sections
Acaules, Acuminatae, Anomalae, Hirsutae, and Lep-
torhizae. The ‘‘core’’ Arctotis group (XIII) consists
of species Lewin placed in the sections Aequales,
Angustifoliae, Argenteae, Asperae, Cuneatae, and
Revolutae. The cypselae in this group vary in terms
of shape, pubescence, abaxial wing morphology,
and, to a lesser extent, pappus and coma length.
Other morphological features and molecular data
are required to further elucidate relationships with-
in this group. It is important to note that Lewin’s
sections were based on involucral bract and vege-
tative morphology, as well as cypselar morphology.
Cypselae of A. verbascifolia (XI), placed in sec-
tion Adpressae by Lewin, have a close affinity with
those of A. acaulis. Lewin distinguished section Ad-
pressae by its absence of appendages on the outer
involucral bracts. That character may vary among
closely related taxa, as in the A. arctotoides species
complex. Both sections Acaules and Adpressae are
acaulescent or very shortly caulescent.
Arctotis graminea is distinctive and is known
only from the Riversdale district, Western Cape
province in South Africa. Lewin (1922a) placed it
in section Revolutae and acknowledged not seeing
mature cypselae when describing the species. Cyp-
selar morphology indicates A. graminea is not close
to A. cuprea and is misplaced in section Revolutae;
its affinities are not clear at present.
Arctotis discolor and A. perfoliata (VIII) were
placed in A. sect. Austro-orientales by Lewin
(1922a). The cypselae differ from those of other
species in having two well-developed abaxial wings
and two cavities, rather than more weakly devel-
oped wings and a single cavity, suggesting an affin-
ity with ‘‘core’’ Arctotis species (XIII). In overall
morphology, A. perfoliata (including A. discolor)
clearly has an affinity with the other species in sec-
tion Austro-orientales, but cypselar morphology in-
dicates it is somewhat isolated in the section.
For convenience, Arctotis fastuosa is herein
grouped with species of A. sect. Austro-orientales
and Cymbonotus (VII). However, at least three cyp-
sela forms are distinguishable in A. fastuosa, and
the resemblance of the cypselae may simply reflect
convergence rather than a common ancestry. In two
forms of A. fastuosa the cypselae bear very reduced
pappus scales. In Cymbonotus and A. sect. Austro-
orientales no specimen examined possesses cypse-
lae with even vestigial pappus scales. However, the
pappus could easily have been lost during the evo-
lution of these groups. In one of these forms of A.
fastuosa the cypselae have a concave abaxial sur-
face and strongly convex adaxial surface, as in A.
hispidula and Cymbonotus lawsonianus; the cyp-

Volume 92, Number 4
2005
589McKenzie et al.
Morphology of Cypselae
selae are constricted and acute at the base and bear
very reduced pappus scales in a single whorl. The
latter characteristics are suggestive of A. hirsuta
and present an alternative hypothesis of affinity. A
third form of A. fastuosa has cypselae with very
weakly developed abaxial wings and lacks a pap-
pus. Arctotis fastuosa was previously placed in Ven-
idium together with A. hirsuta, A. semipapposa, the
species comprising A. sect. Austro-orientales, and
taxa now included in these species (Lessing, 1832;
Harvey, 1865; Stapf, 1928); after merging Arctotis
and Venidium, Lewin (1922a) placed A. fastuosa in
section Caudatae and A. hirsuta in section Hirsu-
tae.
Cymbonotus
Possession of abaxial wings and collateral cavi-
ties on the cypsela is considered a diagnostic char-
acteristic of Arctotis (Bremer, 1994). However, a
close affinity between the cypselae of Arctotis sect.
Austro-orientales (excluding A. perfoliata) and Cym-
bonotus is immediately apparent. On the basis of
cypselar morphology, the two are inseparable and
are assigned to Group VII. The lateral abaxial ribs
are only moderately developed and strongly in-
flexed and form a single, linear or linear-sinuate
cavity, the cypsela surface is often rugose with uni-
cellular or uniseriate trichomes, the cypselae lack
any trace of a pappus, a coma of twin hairs is ab-
sent, and glandular biseriate trichomes are often
present on the carpopodium. Cymbonotus was pub-
lished by Cassini (1825) and has been distin-
guished from Arctotis by the more weakly developed
abaxial cypsela wings, absence of a pappus and a
coma, and papillose filaments. Venidium was dis-
criminated from Arctotis by the glabrous cypselae
with weakly developed abaxial wings and devoid of
a pappus (Lessing, 1831). Lessing (1832) accepted
eight species in Venidium, but commented in a foot-
note (p. 29), ‘‘Nnm [sic] a praecedente [Cymbono-
tus] satis diversum?’’ Beauverd (1915) reduced
Cymbonotus and Venidium to subgenera of Arctotis
based on the similarities in cypselar morphology.
Lewin (1922a) recognized Cymbonotus, but ac-
knowledged that support from morphological evi-
dence was weak and the main basis for its segre-
gation was its geographical isolation. After seeing
Beauverd’s (1915) treatment, Lewin (1922b) con-
curred with Beauverd’s classification of Cymbono-
tus. Subsequent authors (e.g., Norlindh, 1977; Mur-
ray, 1992; Bremer, 1994) continued to recognize
Cymbonotus without comment on the close morpho-
logical similarity with species currently placed in
Arctotis sect. Austro-orientales. As Arctotis is now
circumscribed, cypselar morphology is clearly in-
sufficient to support the independent segregation of
taxa in Cymbonotus and A. sect. Austro-orientales.
Filament ornamentation has also been used to sup-
port segregation of Cymbonotus, but on its own it
may be an inadequate generic criterion in the Arc-
totidinae.
Arctotis maidenii was treated as Cymbonotus sp.
‘A’ by Murray (1992). Cypselar morphology impli-
cates A. maidenii has a close affinity with A. per-
foliata (both species placed in our Group VIII)
rather than with either Cymbonotus species (placed
in Group VII). However, the vegetative morphology
of A. perfoliata and A. maidenii is quite different.
Arctotis perfoliata is a sprawling decumbent peren-
nial with ovate, subrotund or subcordate leaves 5–
6.3 cm long (Harvey, 1865). Arctotis maidenii is an
annual, multi-stemmed herb with lanceolate, pin-
natifid leaves to 30 cm long (Beauverd, 1915; Mur-
ray, 1992). Additional sources of evidence are
needed to further clarify the taxonomic affinities of
A. maidenii.
Haplocarpha
Circumscription of Haplocarpha has been the
subject of much contention. Cypselar morphology
indicates it is an unnatural genus comprising three
species groups (I, II, and III). These species groups
partially support Beauverd’s (1915) subgeneric
groupings. Our Group I corresponds to H. subg.
Landtia with the addition of H. hastata and exclu-
sion of H. schimperi. Group II corresponds to H.
subg. Haplocarpha with the addition of the recently
transferred H. oocephala (Beyers, 2000). Haplocar-
pha schimperi (Group III) is indicated to be isolated
in Haplocarpha.
Previously, three Haplocarpha species (H. nerv-
osa, H. rueppellii, and H. schimperi) have been seg-
regated in Landtia (Lessing, 1832; Harvey, 1865;
Bentham, 1873a, b), but attempts to characterize
Landtia by morphological criteria have proved un-
satisfactory. Haplocarpha was published by Lessing
(1831) and distinguished from other Arctotidinae
by the presence of a uniseriate pappus of scales,
scabrous filaments, pistillate ray florets, and ab-
sence of cypselar wings. Lessing (1832) distin-
guished Landtia from Haplocarpha primarily by its
smooth filaments. However, the filaments are
smooth in H. parvifolia (Lewin, 1922a). Additional
characters discriminating Haplocarpha and Landtia
have been proposed, such as pappus scale apex
shape, cypsela shape, rugose versus smooth cypsela
surface, pappus length relative to cypsela length,
and lengths of peduncles relative to leaves (e.g.,

590 Annals of the
Missouri Botanical Garden
Harvey, 1865; Bentham, 1873a, b; Phillips, 1951).
Closer scrutiny found them to be unsatisfactory and
Landtia was merged with Haplocarpha (Beauverd,
1915; Lewin, 1922a). Beauverd (1915) concluded
the distinction of pappus length relative to cypsela
length and the presence of a median ‘‘vein’’ on the
pappus scales were sufficient to distinguish Landtia
as a subgenus of Haplocarpha.
Chromosome number and cypsela anatomy in-
dicate H. rueppellii (I) has only a distant affinity
with other Arctotidinae. Haplocarpha rueppellii has
2n
5
30 (Norlindh, 1977), in contrast to all other
known chromosome counts in the Arctotidinae (in-
cluding counts for H. lyrata and H. scaposa)of2n
5
18 (e.g., Norlindh, 1977; Turner, 1970; Ober-
prieler & Vogt, 1993; Strother et al., 1996; Carr &
King, 1999; Skinner, 2002). In this regard, H. ruep-
pellii may have a distant affinity with Gazania and
Hirpicium Cass. of the Gorteriinae, for which counts
of 2n
5
10 and 2n
5
20 have been reported (Nor-
lindh, 1977). In terms of cypsela anatomy, H. ruep-
pellii possesses a cypsela form shared only with H.
scaposa; H. lyrata cypselae are similar to those of
Arctotheca, Arctotis, and Cymbonotus (Reese, 1989).
Cypselar morphology supports the transfer of H.
oocephala (II) from Arctotis by Beyers (2000). The
cypselae of H. oocephala are very similar to those
of H. parvifolia, which is classified in H. subg. Hap-
locarpha (Beauverd, 1915), and lack the abaxial
wings and cavities characteristic of Arctotis. Hap-
locarpha subg. Haplocarpha (including H. oocepha-
la) forms a coherent group (II) characterized by
cypselae with well-developed abaxial ridges, a pap-
pus of free scales longer than the cypsela, and a
coma of twin hairs at least equal in length to the
cypsela. Distinctive features restricted to certain
species in this group include shortly inflexed lateral
ridges in H. lyrata, extremely broad abaxial ridges
in H. oocephala and H. parvifolia, and acuminate
pappus scale apices in H. lyrata and H. scaposa.
The trichome type present on the cypsela surface
varies in this group; long twin hairs exceeding the
cypsela length cover the surface in H. lanata and
H. lyrata, biseriate glandular trichomes are present
on the abaxial ridges in H. oocephala and H. par-
vifolia, and the cypsela surface is glabrous in H.
scaposa.
Cypselar morphology indicates H. schimperi (III)
is not closely related to other species currently
placed in Haplocarpha. The generic affinities of H.
schimperi have long been uncertain. Haplocarpha
schimperi, along with H. rueppellii, initially was
placed in the specially erected genus Schnittspahn-
ia Sch. Bip. (Schultz Bipontinus, 1842). Recogniz-
ing the morphological incongruence between the
two species, Gay (1847) transferred S. schimperi
Sch. Bip. to a new genus, Ubiaea J. Gay, before
Bentham (1873a, b) merged both Schnittspahnia
and Ubiaea with Landtia. Following the reduction
of Landtia to a subgenus within Haplocarpha
(Beauverd, 1915), H. schimperi still does not sit
well in Haplocarpha. Haplocarpha schimperi is sim-
ilar to subgenus Landtia in possessing a reduced
pappus and lacking a coma, but differs in possess-
ing well-developed, dentate abaxial ribs. From sub-
genus Haplocarpha, H. schimperi is distinguished
by the reduced, sometimes coronate pappus, ab-
sence of a coma, and presence of uniseriate tri-
chomes on the cypselae. Cypselar morphology sug-
gests some affinity with Arctotis fastuosa and
Cymbonotus preissianus, but H. schimperi differs
from those species in consistently possessing a pap-
pus and the abaxial ribs are not developed into
wings.
Presence of well-developed abaxial ridges or
wings on the cypsela is a possible synapomorphy
of Arctotidinae (Funk et al., 2004). The weak de-
velopment of the abaxial ribs in our Group I casts
some doubt on this and indicates coding of this
character requires some caution. Absence of con-
spicuous abaxial ridges or wings may be plesiom-
orphic in Arctotideae. Alternatively, a concomitant
reduction of the abaxial ridges and pappus in
Group I may be autapomorphic and would be con-
sistent with the possession of geocarpic capitula in
H. nervosa and H. rueppellii (Barker, 2005 this is-
sue).
It is clear that a thorough reappraisal of other
morphological characters and acquisition of molec-
ular data is required to elucidate relationships
among all Haplocarpha species and to clarify the
generic circumscription.
C
ONCLUSIONS
This study confirms the utility of cypselar mor-
phology for elucidating taxonomic relationships in
Arctotidinae. Arctotis is indicated to comprise a
large core group containing 11 of the species sam-
pled; the other 14 species sampled are placed in
eight different species groups. Lewin’s infrageneric
classification of Arctotis is partly supported, but ad-
ditional morphological and molecular evidence is
required to further resolve infrageneric relation-
ships. Arctotis sect. Austro-orientales (excluding A.
perfoliata) and Cymbonotus are indicated to have a
close affinity, and an expanded Cymbonotus might
be warranted. Haplocarpha is indicated to be an
unnatural genus as circumscribed at present. Res-
urrection of Landtia is supported by cypselar mor-

Volume 92, Number 4
2005
591McKenzie et al.
Morphology of Cypselae
phology. The affinities of H. schimperi are less cer-
tain; it seems misplaced in Haplocarpha. Arctotheca
comprises a coherent group, but resolution of the
circumscription of A. calendula and A. prostrata is
required. Corroboration with other morphological
characters and molecular data is still required for
resolution of taxonomic relationships in Arctotidi-
nae.
This study also highlights the importance of in-
cluding as wide a sample of taxa as possible in
initial investigations of taxonomic relationships. Se-
lection of a too small sample of randomly selected
representatives of genera, in the absence of knowl-
edge of species groupings, may generate misleading
relationships and conclusions, particularly regard-
ing monophyly and delimitation of supraspecific
taxa. Future investigations of Arctotidinae should
include one or more representatives of Haplocarpha
subg. Landtia, especially when studying the mono-
phyly of the subtribe. The present study indicates
five species (Arctotis breviscapa, A. dregei, A. mai-
denii, A. perfoliata, and Haplocarpha schimperi) are
anomalous as classified at present and should be
included in future phylogenetic studies to elucidate
their relationships.
Literature Cited
Barker, N. P. 2005. A review and survey of basicarpy,
geocarpy, and amphicarpy in the African and Madagas-
can flora. Ann. Missouri Bot. Gard. 92: 445–462.
Beauverd, G. 1915. Contribution a` l’e´tude des Compo-
se´es: suite X. Bull. Soc. Bot. Gene`ve (se´r. 2) 7: 21–56.
Beneke, K., I. von Teichman, M. W. van Rooyen & G. K.
Theron. 1992. Fruit polymorphism in ephemeral spe-
cies of Namaqualand: II. Anatomical differences be-
tween polymorphic diaspores of Arctotis fastuosa and
Ursinia cakilefolia. S. African J. Bot. 58: 456–460.
Bentham, G. 1873a. Compositae. Pp. 1–554 in G. Ben-
tham & J. D. Hooker (editors), Genera plantarum ad
exemplaria imprimis in herbariis kewensibus servata
definata, Vol. 2. L. Reeve, London.
. 1873b. Notes on the classification, history, and
geographical distribution of Compositae. J. Linn. Soc.,
Bot. 13: 335–582.
Beyers, J. B. P. 2000. Haplocarpha. P. 329 in P. Goldblatt
& J. C. Manning, Cape Plants: A Conspectus of the
Cape Flora of South Africa. Strelitzia 9. National Bo-
tanical Institute, Pretoria, and Missouri Botanical Gar-
den, St. Louis.
. 2002. A new species of Arctotheca from North-
ern Cape, South Africa. Bothalia 32: 185–187.
Breitwieser, I. & J. M. Ward. 2003. Phylogenetic rela-
tionships and character evolution in New Zealand and
selected Australian Gnaphalieae (Compositae) inferred
from morphological and anatomical data. Bot. J. Linn.
Soc. 141: 183–203.
Bremer, K. 1994. Asteraceae: Cladistics and Classifica-
tion. Timber Press, Portland.
Candolle, A. P. de. 1837. Prodromus systematis naturalis
regni vegetabilis, sive enumeratio contracta ordinum,
generum, specierumque plantarum huc usque cognitar-
um, juxta methodi naturalis normas digesta, Vol. 6.
Treuttel & Wu¨ rtz, Paris.
Carr, G. D. & R. M. King. 1999. Chromosome numbers
in Compositae. XVIII. Amer. J. Bot. 86: 1003–1014.
Cassini, H. 1825. Odontoptera. Pp. 396–398 in G. L.
Cuvier (editor), Dictionnaire des Sciences Naturelles,
Vol. 35. Le Normant, Paris.
Compton, R. H. 1953. Plantae novae africanae. Ex Africa
semper aliquid novi.–Pliny. Series XXXI. J. S. African
Bot. 19: 109–134.
Cronquist, A. 1985. History of generic concepts in the
Compositae. Taxon 34: 6–10.
Funk, V. A., R. Chan & S. C. Keeley. 2004. Insights into
the evolution of the tribe Arctoteae (Compositae: sub-
family Cichorioideae s.s) using trnL-F, ndhF, and ITS.
Taxon 53: 637–655.
Gay, J. 1847. Ubiaea. P. 447 in A. Richard (editor), Ten-
tamen florae abyssinicae seu enumeratio plantarum
hucusque in plerisque Abyssiniae provinciis detectar-
um et praecipue a beatis doctoribus Richard Quartin
Dillon et Antonio Petit (annis 1838–1843) lectarum,
Vol. 1. Arthus Bertrand, Paris.
Germishuizen, G. & N. L. Meyer (editors). 2003. Plants
of Southern Africa: An Annotated Checklist. Strelitzia
14. National Botanical Institute, Pretoria.
Goldblatt, P. & J. C. Manning. 2000. Cape Plants: A
Conspectus of the Cape Flora of South Africa. Strelitzia
9. National Botanical Institute, Pretoria, and Missouri
Botanical Garden, St. Louis.
Harvey, W. H. 1865. Compositae. Pp. 44–530 in W. H .
Harvey & O. Sonder (editors), Flora Capensis, Vol. 3.
L. Reeve, London.
Herman, P. P. J. 2001. Observations on hairs in the ca-
pitula of some southern African Asteraceae genera. S.
African J. Bot. 67: 65–68.
Hess, R. 1938. Vergleichende Untersuchungen u¨ber die
Zwillingshaare der Compositen. Bot. Jahrb. Syst. 68:
435–496.
Hilliard, O. M. 1977. Compositae in Natal. Univ. Natal
Press, Pietermaritzburg.
Karis, P. O., P. Eldena¨s&M.Ka¨llersjo¨. 2001. New evi-
dence for the systematic position of Gundelia L. with
notes on delimitation of Arctotideae (Asteraceae). Taxon
50: 105–114.
Leins, P. 1970. Die Pollenko¨rner und Verwandtschafts-
beziehungen der Gattung Eremothamnus (Asteraceae).
Mitt. Bot. Staatssamml. Mu¨ nchen 7: 369–376.
Lessing, C. F. 1831. Synanthereae Rich. Pp. 83–170 in
L. C. A. von Chamisso & D. F. L. von Schlechtendal,
De plantis in expeditione speculatoria Romanzoffiana
observatis. Linnaea 6: 76–170.
. 1832. Synopsis generum compositarum ear-
umque dispositionis novae tentamen, monographiis
multarum capensium interjectis. Duncker & Humblot,
Berlin.
Lewin, K. 1922a. Systematische Gliederung und geogra-
phische Verbreitung der Arctotideae–Arctotidinae. Re-
pert. Spec. Nov. Regni Veg. Beih. 11: 1–75.
. 1922b. Zu meiner Arbeit u¨ ber die Arctotidinae.
Repert. Spec. Nov. Regni Veg. 18: 300.
Murray, L. J. 1992. Cymbonotus. P. 318 in G. J. Harden
(editor), Flora of New South Wales, Vol. 3. NSW Univ.
Press, Kensington.
Norlindh, T. 1977. Arctoteae—Systematic review. Pp.
943–959 in V. H. Heywood, J. B. Harborne & B. L.
Turner (editors), The Biology and Chemistry of the Com-
positae, Vol. 2. Academic Press, London.

592 Annals of the
Missouri Botanical Garden
Oberprieler, C. & R. Vogt. 1993. Chromosome numbers
of north African phanerogams. II. Willdenowia 23: 211–
238.
Phillips, E. P. 1951. The Genera of South African Flow-
ering Plants. Department of Agriculture, Pretoria.
Reese, H. 1989. Die Entwicklung von Pericarp und Testa
bei Calenduleae und Arctotideae (Asteraceae)—ein
Beitrag zur Systematik. Bot. Jahrb. Syst. 110: 325–419.
Rourke, J. P. 1974. Dymondia margaretae, a notable
ground cover. J. Bot. Soc. S. Africa 60: 30–33.
Schultz Bipontinus, C. H. 1842. U
¨
ber die Compositae
von Eduard Ru¨ppell’s und Wilhelm Schimper’s abyssin-
ischer und Kotschy’s nubischer Reise. Flora 25: 433–
448.
Skinner, A. 2002. The Use of Morphological and DNA
Sequencing Data to Determine Generic Relationships
of the Sub-tribe Arctotidinae (Arctotideae Asteraceae).
Unpublished B.Sc. (Hons) thesis, Rhodes University,
South Africa.
Stapf, O. 1928. Venidium fastuosum. Bot. Mag. t. 9127.
Strother, J. L., L. E. Watson & J. L. Panero. 1996. Doc-
umented chromosome numbers 1996.3. Chromosome
numbers in some South African Compositae. Sida 17:
265–268.
Turner, B. L. 1970. Chromosome numbers in the Com-
positae. XII. Australian species. Amer. J. Bot. 57: 382–
389.
A
PPENDIX
1. C
YPSELAR
M
ORPHOLOGICAL
C
HARACTERS
FOR
A
RCTOTIDINAE IN
T
ABLE
2.
1. Cypsela shape: (1) obconical; (2) obovoid to obovoid-
cylindrical; (3) obovoid-elliptic; (4) obconical-ob-
ovoid; (5) obconical-cylindrical.
2. Cypsela bilaterally flattened: (0) absent; (1) present.
3. Cypsela base constricted forming a short ‘‘stalk’’: (0)
absent; (1) present.
4. Cypsela epidermis transversely rugose: (0) absent; (1)
abaxial surface/wings; (2) adaxial surface; (3) both
surfaces.
5. Median abaxial rib width: (0) not very conspicuous;
(1) narrow; (2) broad.
6. Median abaxial rib transversely rugose: (0) absent; (1)
present.
7. Median abaxial rib length v. lateral abaxial rib
length: (1) median rib shorter; (2)
6
equal length; (3)
median rib longer.
8. Origin of cypsela wings: (0) wings absent; (1) abaxial
ribs only; (2) adaxial and abaxial ribs.
9. Number of cavities formed by abaxial ribs.
10. Number of cavities formed by adaxial ribs.
11. Shape of abaxial cavity: (1) obovoid; (2) linear; (3)
linear-cuneiform; (4) triangular-pandurate; (5) pan-
durate-linear; (6) linear-ovate; (7) ovate.
12. Cypsela wing margin: (1) smooth; (2) dentate.
13. Cypsela wing teeth apex: (1) acute; (2) obtuse; (3)
rounded; (4) truncate.
14. Abaxial wing curvature toward median abaxial rib: (1)
weakly incurved; (2) moderately incurved; (3) abaxial
cavities
6
occluded by wings.
15. Trichome type on cypsela epidermis: (1) septate tri-
chomes; (2) uniseriate trichomes with rounded apices;
(3) twin hairs; (4) biseriate glandular trichomes; (5)
narrow unicellular trichomes with acute apices; (6)
clavate uniseriate trichomes.
16. Trichome density on abaxial surface: (0) glabrous; (1)
sparse; (2) moderate; (3) dense.
17. Trichome density on adaxial surface: (0) glabrous; (1)
sparse; (2) moderate; (3) dense.
18. Trichomes on carpopodium: (0) absent; (1) complete
ring; (2) tuft/s on abaxial side only; (3) partial ring on
adaxial side only.
19. Maximum coma length v. cypsela length.
20. Coma trichome type: (1) twin hairs; (2) multicelluar
biseriate trichomes; (3) biseriate glandular trichomes.
21. Trichomes on apical plate: (0) absent; (1) present.
22. Number of pappus series.
23. Number of pappus scales in outer series.
24. Number of pappus scales in inner series.
25. Relative length of pappus series: (1)
6
equal; (2) inner
series at least double outer series.
26. Pappus length v. cypsela length.
27. Pappus scale fusion: (0) all scales free; (1) coronate.
28. Pappus scale base: (1) narrow; (2) broad.
29. Pappus scale apex shape: (0) acuminate; (1) acute; (2)
obtuse; (3) rounded.

Volume 92, Number 4
2005
593McKenzie et al.
Morphology of Cypselae
A
PPENDIX
2. Voucher details of specimens examined. The abbreviation of the herbarium where each specimen is
lodged is in parentheses. Geographic abbreviations: EC
5
Eastern Cape, South Africa; ETH
5
Ethiopia; FS
5
Free
State, South Africa; KZN
5
KwaZulu-Natal, South Africa; LES
5
Lesotho; NAM
5
Namibia; NC
5
Northern Cape,
South Africa; NSW
5
New South Wales, Australia; SA
5
South Australia, Australia; TAS
5
Tasmania, Australia; VIC
5
Victoria, Australia; WC
5
Western Cape, South Africa. Collector abbreviations: JS
5
J. Samuel; NPB
5
N. P.
Barker; RJM
5
R. J. McKenzie. Herbarium abbreviations: BOL
5
Bolus Herbarium; ETH
5
National Herbarium of
Ethiopia; GRA
5
Selmar Scho¨nland Herbarium; NBG
5
Compton Herbarium; NH
5
KwaZulu-Natal Herbarium; NSW
5
National Herbarium of New South Wales; PRE
5
Pretoria National Herbarium.
Taxa Voucher specimens
Arctotheca calendula (L.) Levyns EC, Convent 40 (GRA); KZN, C. J. Ward 5826 (PRE); KZN, C. L. P. Zeyher
s.n. (GRA); WC, C. Cottrell 7 (GRA); WC, M. Francis 11 (GRA); WC, D.
W. Gess 91/92/3 (GRA); WC, S. K. Gess, F. W. Gess & D. W. Gess 89/90/38
(GRA); WC, JS 16 (GRA); WC, JS 19 (GRA); WC, JS 37 (GRA); WC, JS
45 (GRA); WC, JS 48 (GRA); WC, RJM 808/1 (GRA); WC, RJM 1069/2
(GRA); WC, R. Schlechter 8123 (GRA); WC, W. C. Scully 30 (BOL).
Arctotheca calendula (L.) Levyns
3
Arctotheca populifolia (P. J. Ber-
gius) Norl.
WC, C. Peter 84 (GRA).
Arctotheca forbesiana (DC.) K.
Lewin
WC, E. E. Esterhuysen 23126 (PRE).
Arctotheca populifolia (P. J. Bergi-
us) Norl.
EC, D. M. Comins 1041 (GRA); EC, C. J. Cooper 32 (GRA); EC, O. Hilmer
187 (GRA); EC, C. Peter 211 (GRA); EC, C. Peter 258 (GRA); EC, G. Rat-
tray 579 (GRA); WC, RJM 771 (GRA); WC, C. Peter 287 (GRA).
Arctotheca prostrata (Salisb.) Brit-
ten
EC, H. H. Burrows 4679 (GRA); EC, R. M. Cowling 338 (GRA); EC, RJM
1129 (GRA); WC, P. A. Bean 2992 (BOL); WC, RJM 828 (GRA); WC,
RJM 829 (GRA); WC, S. Scho¨nland 3516 (GRA); WC, A. J. Williamson 76
(GRA).
Arctotis acaulis L. WC, A. Hitchcock 924 (GRA); WC, C. P. Immelman 1866/28 (BOL); WC,
RJM 848/1 (GRA); WC, RJM 1030/3 (GRA); WC, RJM 1041/3 (GRA);
WC, RJM 1047 (GRA); WC, W. G. Welman 106 (PRE).
Arctotis angustifolia L. WC, L. Bolus s.n. (BOL); WC, E. E. Esterhuysen 31373 (NBG).
Arctotis arctotoides (L. f.) O. Hoffm. EC, RJM 724/1 (GRA); EC, RJM 778/4 (GRA); EC, RJM 781/2 (GRA); EC,
RJM 790/1 (GRA); EC, RJM 793/1 (GRA); EC, RJM 884/2 (GRA); EC,
RJM 887 (GRA); EC, RJM 905/1 (GRA); EC, RJM 908 (GRA); EC, RJM
969/3 (GRA); FS, Gray College Herb. 107 (BOL); KZN, M. S. Evans 427
(NH); LES, M. M. Page 15979 (BOL); LES, M. D. Panagos 26 (NH).
Arctotis argentea Thunb. WC, E. E. Esterhuysen 29344 (BOL); WC, E. E. Esterhuysen 29347 (BOL);
WC, H. C. Taylor 12022 (NBG); WC, H. C. Taylor 1210, (NBG); WC, E.
van Jaarsveld 12108 (PRE).
Arctotis aspera L. var. aspera WC, RJM 847/1 (GRA); WC, RJM 895/1 (GRA).
Arctotis breviscapa Thunb. WC, C. Boucher & P. Shepherd 4602 (NBG, PRE); WC, J. A. Marsh 1214
(NBG); WC, F. A. Rogers 1062 (GRA); WC, E. Werderman & H.-D. Ober-
dieck 473 (PRE).
Arctotis campanulata DC. WC, M. Lewis 19410 (BOL).
Arctotis cuneata DC. EC, NPB 1921 (GRA); EC, R. A. Dyer 1686 (GRA); EC, RJM 826/1 (GRA);
EC, M. J. Wells 2937 (GRA).
Arctotis cuprea Jacq. WC, R. Schlechter 8697 (BOL); WC, W. G. Welman 29 (PRE).
Arctotis dregei Turcz. EC, H. H. Burrows 4639 (GRA); EC, J. Burtt Davy 14245A (BOL); EC, RJM
1016/1 (GRA); EC, P. Raal 1830 (GRA); NC, J. Hutchinson 690 (BOL);
WC, K. A
˚
. Dahlstrand 1445 (PRE); WC, RJM 815/2 (GRA); WC, RJM
833/2 (GRA); WC, RJM 834/1 (GRA); WC, RJM 1050/2 (GRA).
Arctotis fastuosa Jacq. NAM, Dinter 6063 (BOL); NAM, H. Merxmu¨ller & W. Giess 28462 (PRE);
NC, P. Burgoyne 1517 (PRE); NC, R. H. Compton 11488 (NBG); NC, B.
Maguire 1907 (NBG); NC, D. C. H. Plowes 7274 (PRE); NC, M. F. Thomp-
son 1066 (PRE); NC, M. F. Thompson & A. le Roux 211 (NBG); WC, S. M.
Johnson 540 (NBG); WC, H. H. W. Pearson 6558 (BOL).
Arctotis graminea K. Lewin WC, RJM 824/2 (GRA); WC, RJM 1084/2 (GRA).
Arctotis gumbletonii Hook. f. WC, J. W. Mathews s.n., x.1931 (BOL).

594 Annals of the
Missouri Botanical Garden
A
PPENDIX
2. Continued.
Taxa Voucher specimens
Arctotis hirsuta (Harv.) Beauverd EC, RJM 766, ex cult. (GRA); WC, RJM 1070/2 (GRA); WC, N. S. Pillans
7693 (BOL); WC, N. S. Pillans 8628 (BOL).
Arctotis hispidula (Less.) Beauverd EC, J. B. Gillett 2350 (NBG); EC, RJM 760/1 (GRA).
Arctotis maidenii Beauverd NSW, W. E. Mulham s.n., viii.1971 (NSW); NSW, W. E. Mulham 1145 (NSW);
NSW, T. & J. Whaite 3787 (NSW).
Arctotis microcephala (DC.) Beau-
verd
WC, W. J. Hanekom 1963 (PRE).
Arctotis perfoliata (L. f.) O. Hoffm. EC, W. E. Dix 156 (NBG); WC, E. E. Esterhuysen 31531 (BOL); WC, H. G.
Fourcade 64 (BOL); WC, RJM 825/2 (GRA).
Arctotis rotundifolia K. Lewin WC, H. H. Bolus 1192 (BOL); WC, E. E. Esterhuysen 11129 (BOL); WC, R.
Schlechter 9249 (BOL).
Arctotis scapiformis Thell. EC, A. White 41 (GRA); KZN, T. A. Coleman 475 (NH).
Arctotis scullyi Du¨mmer NC, E. J. van Jaarsveld 7423, (NBG); WC, H. H. Bolus 9615, (BOL).
Arctotis semipapposa (DC.) Beau-
verd
WC, Gentry, Barclay & Van Breda´ 18626 (PRE); WC, RJM 890 (GRA); WC,
RJM 1178 (GRA).
Arctotis suffruticosa K. Lewin EC, R. Schlechter 2586 (BOL, GRA [isotypes]).
Arctotis venusta Norl. EC, J. E. Victor 1844 (GRA); FS, H. H. Bolus s.n., ii.1904 (BOL); FS, RJM
873/1 (GRA); FS, RJM 878/1 (GRA); WC, Kitto s.n., ex cult., x.1931
(BOL).
Arctotis verbascifolia Harv. WC, E. G. H. Oliver 3951 (NBG).
Cymbonotus lawsonianus Gaudich. NSW, J. L. Boorman s.n., xii.1916 (NSW); NSW, R. Coveny 8772 (NSW);
NSW, R. Coveny 10217 (NSW); NSW, R. Coveny 7406, D. Benson & H.
Bryant (NSW); NSW, J. M. Dalby 86/128 (NSW); NSW, G. L. Davis s.n.,
16.1.1941 (NSW); NSW, K. Mair s.n., 20.x.1957 (NSW); NSW, E. J. Mc-
Barron 14438 (NSW); NSW, S. McIntyre 5-8 (NSW); NSW, J. Pickard 296
& D. Blaxell (NSW).
Cymbonotus preissianus Steetz NSW, E. F. Constable s.n., 25.x.1958 (NSW); NSW, R. Coveny 10773 & J.
Armstrong (NSW); NSW, A. Rodd 1606 (NSW); NSW, H. Salasoo 2113
(NSW); SA, D. N. Kraehenbuehl 4835 (PRE); TAS, R. C. Gunn 509 (NSW);
VIC, S. G. Hannaford s.n., viii.1853 (NSW); VIC, J. R. Hosking 49 (NSW).
Haplocarpha hastata K. Lewin ETH, J. W. Ash 1700 (ETH); ETH, O. Hedberg 5698 (ETH).
Haplocarpha lanata Less. WC, F. J. Kruger 441 (PRE).
Haplocarpha lyrata Harv. EC, RJM 1116 (GRA).
Haplocarpha nervosa (Thunb.)
Beauverd
EC, R. A. Dyer 786 (GRA); EC, RJM 1125/1 (GRA).
Haplocarpha oocephala (DC.)
Beyers
WC, E. E. Esterhuysen 34661 (BOL).
Haplocarpha parvifolia (Schltr.)
Beauverd
WC, E. E. Esterhuysen 36048 (PRE).
Haplocarpha rueppellii (Sch. Bip.)
Beauverd
ETH, M. Thulin, A. Hunde & M. Tadesse 3715 (ETH).
Haplocarpha scaposa Harv. EC, RJM 909/4 (GRA).
Haplocarpha schimperi (Sch. Bip.)
Beauverd
ETH, NPB 1899 (GRA, ETH); ETH, S. H/Manam 1 (ETH); ETH, M. Thulin
1609 (ETH).