ArticlePDF Available

Abstract and Figures

The relationships between Hemizygia and Syncolostemon are investigated. Parsimony analyses of morphological data and sequences of the trnL intron, trnL-trnF intergenic spacer and the rps16 intron were carried out separately and in combination. All analyses largely show strong support for the two genera as a monophyletic group; neither genus as currently circumscribed is monophyletic. They are herein merged under the earliestname Syncolostemon. The clade can be diagnosed by the presence of fused anterior stamens. Bremer’s ancestral area analysis method suggests that the Afroalpine and Afromontane archipelago-like Regional Centre of Endemism may be the ancestral area for Syncolostemon s.l.
Content may be subject to copyright.
INTRODUCTION
The genera Hemizygia (Benth.) Briq. and Syncolos-
temon E. Mey. ex Benth. belong in the tribe Ocimeae of
the Labiatae (Cantino & al., 1992; Paton, 1998; Paton &
al., 2004), a large natural tribe of over 1000 species that
is almost exclusively tropical. The tribe is characterized
by the presence of declinate stamens and synthecous
anthers (Paton, 1998). However, this stamen character is
not always consistent, as some members of the tribe, for
e.g., Tetradenia Benth. (Wagstaff & al., 1995) and certain
elements of Platostoma sect. Limniboza (R.E. Fr.) A.J.
Paton, have spreading stamens (Paton, 1997). Hemizygia
and Syncolostemon have fused anterior stamens, a puta-
tive synapomorphy, that sets them apart from other gen-
era in tribe Ocimeae (Paton, 1998). Hemizygia compris-
es perennial soft shrubs or annual herbs and Syncoloste-
mon perennial herbs or soft shrubs. Currently there are
33 species recognized in Hemizygia including the two
species, Hemizygia madagascariensis A.J. Paton &
Hedge (Paton & Hedge, 1999) and H. comosa (Wright ex
Benth.) A.J. Paton (Paton, 1998) from Madagascar and
India respectively. Apart from H. madagascariensis and
H. comosa, all other species in the genus occur in main-
land Africa and among these, only H. bracteosa (Benth.)
Briq. and Hemizygia welwitschii (Rolfe) M. Ashby
extend beyond the southern Africa region. Syncoloste-
mon comprises 11 species including the reinstated Syn-
colostemon ramulosus E. Mey. ex Benth. (Codd, 1988)
and the new combination, S. flabellifolius (S. Moore)
A.J. Paton (Paton, 1998) from the Chimanimani
Mountains of northeastern Zimbabwe. Both Hemizygia
and Syncolostemon occur in similar habitats, usually dry,
often rocky, woodland or grassland though there are
some species, e.g., H. stalmansii A.J. Paton & K. Balk-
will, which also occur on serpentine but not exclusively
(Paton & Balkwill, 2001).
Bentham (1837) first described Syncolostemon in
which he accommodated three species. Later, the genus
was divided into two sections: sect. Micranthi and sect.
Macranthi (Briquet, 1897). In a fairly recent taxonomic
account of the genus, two informal groups based on dif-
ferences in calyx, corolla, inflorescence and nutlet char-
acteristics were proposed (Codd, 1976a). Five species
viz. S. parviflorus E. Mey. ex Benth., S. concinnus N.E.
Br., S. argenteus, S. eriocephalus Verdoorn and S. comp-
tonii Codd, characterised by small corollas and calyces,
lax inflorescence and nutlets with a frill-like extension at
the base, were placed in Group 1. The second group
(Group 2) was diagnosed by the presence of a large
corolla and calyx, dense inflorescence and lacked the
frill-like extension at the base of the nutlets. Only two
species, S. rotundifolius E. Mey. ex Benth and S. densi-
florus Benth. were placed in this group.
Hemizygia was first recognized as a section of Ocim-
um L. by Bentham (1848) to accommodate the species,
O. teucriifolium Hochst., which was found to differ from
typical Ocimum in having anterior staminal filaments
fused at the base. However, the section was later raised
to generic level by Briquet (1897). Baker (1900) includ-
ed Hemizygia in Orthosiphon Benth. but gave no reasons
for doing this. Brown (1910) rejected the fusion of the
941
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
A reassessment of Hemizygia and Syncolostemon (Ocimeae—Lamiaceae)
Donald F. Otieno1,2, Kevin Balkwill1, Alan J. Paton3& Vincent Savolainen4
1School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, P/Bag 3, WITS 2050,
Johannesburg, South Africa.
2Present address: Department of Botany, Moi University, P. O. Box 3900, Eldoret, Kenya. dfotieno@yahoo.co
.uk (author for correspondence).
3Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, U.K.
4Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, U.K.
The relationships between Hemizygia and Syncolostemon are investigated. Parsimony analyses of morpholog-
ical data and sequences of the trnL intron, trnL-trnF intergenic spacer and the rps16 intron were carried out
separately and in combination. All analyses largely show strong support for the two genera as a monophyletic
group; neither genus as currently circumscribed is monophyletic. They are herein merged under the earliest
name Syncolostemon. The clade can be diagnosed by the presence of fused anterior stamens. Bremer’s ances-
tral area analysis method suggests that the Afroalpine and Afromontane archipelago-like Regional Centre of
Endemism may be the ancestral area for Syncolostemon s.l.
KEYWORDS:
generic delimitation, Hemizygia, Syncolostemon, phylogenetic analysis.
filaments of the anterior stamens as a generic character in
Hemizygia, stating that it varied considerably in some
species being, for e.g., united in some flowers but also
quite free in others in the same raceme, e.g., in
Orthosiphon persimilis N.E. Br. [= Hemizygia persimil-
is]. He therefore subsumed Hemizygia under Orthosiph-
on Benth. Ashby (1935) reinstated Hemizygia to generic
rank noting that although the degree of union of the ante-
rior filaments in the group varied, it was often not to any
marked degree and only rarely were the filaments free to
the base. Based mainly on floral characters, Ashby
(1935) then distinguished Hemizygia from its two closest
allies, Orthosiphon and Ocimum, but concluded that it
was intermediate between them. He also implied a close
affinity between Hemizygia and Syncolostemon by
observing that H. flabellifolia S. Moore [= S. flabelli-
folius], H. obermeyerae M. Ashby and H. latidens (N.E.
Br.) M. Ashby [= S. latidens (N.E. Br.) Codd] were, “in
general appearance and in the nature of the floral bracts,
suggestive of the genus Syncolostemon”. Ashby’s obser-
vations were confirmed in later studies (e.g., Codd,
1976a, b; Paton, 1998; Paton & Balkwill, 2001), which
reported similarities between the two genera in corolla
and androecial characters. Besides resurrecting Hemizy-
gia, Ashby also subsumed the genus Bouetia A. Chev.
under it, observing that the type of Bouetia was not dis-
tinguishable from H. bracteosa.
The only character that has consistently been used to
separate Hemizygia and Syncolostemon is the shape of
the mature calyx. In Hemizygia the upper calyx lip is
broadly ovate, often decurrent on the calyx tube and
much larger than the lower four teeth which are subulate
to spinescent (see Fig. 4D), while in Syncolostemon the
calyx is more or less equally five-toothed (see Fig. 4A)
(Codd, 1976a, b), the teeth being deltoid to deltoid-sub-
ulate. However, in S. latidens and S. macranthus, the
upper calyx teeth are distinctly longer than the lower four
and they are elliptic to broadly elliptic (Codd, 1976a). On
the basis of these characters the two species have been
considered to be intermediate between Syncolostemon
and Hemizygia and the two genera, hence, regarded as
being continuous with one another (Codd, 1976a). The
similarities in corolla and habit between H. stalmansii
and S. macranthus and the scarcely decurrent young
calyces of the former further weaken the argument for
retaining separate genera (Paton & Balkwill, 2001).
Codd (1976a) advocated merging Hemizygia and
Syncolostemon but refrained from doing so, noting that
since Syncolostemon was the earlier generic name it
would have taken precedence thus necessesitating many
recombinations. In agreement with Codd (1976a), Paton
(1998) also expressed reluctance to synonymise the gen-
era despite the lack of unambiguous synapomorphies,
until all available material was comprehensively studied.
Ryding (1992) placed Hemizygia, Syncolostemon
and Ocimum together with Becium Lindl. [= Ocimum],
Catoferia (Benth.) Benth. and Erythrochlamys Gürke in
his “Ocimum group” of the subtribe Ociminae and in a
parsimony analysis of Ocimum and its close allies, Paton
& al. (1999) showed that there was a close relationship
between Syncolostemon and Hemizygia. The two united
as a monophyletic group and emerged as sister to an
enlarged monophyletic Ocimum but differed from the lat-
ter in having an unequally lobed disk and fused anterior
stamens (Paton & al., 1999). Paton & al. (1999) did not,
however, address the question of whether Syncolostemon
and Hemizygia were congeneric but emphasized the need
for a study to tackle the problem. They recommended
that such a study should be based on the whole range of
the complex including the Madagascan and Indian rela-
tives. In a recent study of the tribe Ocimeae, based on the
trnL-F and rps16 regions of the chloroplast genome and
morphology (Paton & al., 2004), relationships within
Ociminae were not supported by bootstrap values over
50%. However the strict consensus tree supported Oci-
mum as sister to a monophyletic Hemizygia/Syncoloste-
mon clade.
The aims of this study are first, to evaluate, using
both morphological and molecular characters whether
the currently defined Hemizygia and Syncolostemon are
monophyletic; second to determine evolutionary rela-
tionships among their species and third to use the phy-
logeny to investigate their biogeography.
MATERIALS AND METHODS
Sampling of taxa. —
Thirty-three recognized
species of Hemizygia and 11 of Syncolostemon were in-
cluded in the morphological cladistic study. In addition,
two entities considered to be new species of Hemizygia
(included as H. sp. nov.) and Syncolostemon (included as
S. sp. nov.) were studied. The widespread H. bracteosa
was represented by three forms: the typical form (H.
bracteosa 1) diagnosed by the presence of a network of
veins on the nutlet and with a distribution extending from
east and central, to southern Africa; a second form (H.
bracteosa 2) diagnosed by the presence of only a single
vein on the outer surface of the nutlet and only known
from specimens collected in Nigeria, Mali, Benin and
Cameroon; and a third form (H. bracteosa 3) with revo-
lute leaf margins, a parallel-sided corolla tube which
only expands near the mouth, hairs at the insertion posi-
tion of the anterior stamens and also with a single nutlet
vein. This last form, is this far, only known from
Mozambique. Hemizygia ornata S. Moore was included
as a distinct species in agreement with Moore (1911). It
differs from H. welwitschii, under which Ashby (1935)
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
942
subsumed it, in leaf, corolla and gynoecial characters.
Two subspecies of H. pretoriae, subsp. pretoriae (Gürke)
M. Ashby and subsp. heterotricha Codd, were included
separately as H. pretoriae 1 and 2 respectively.
To test whether Hemizygia and Syncolostemon
together form a monophyletic group, this study included
three species from Ocimum: O. tenuiflorum [one of the
basal species in the Ocimum clade in the analyses of
Paton & al. (1999, 2004)], Ocimum serratum (Schltr.)
A.J. Paton and O. labiatum (N.E. Br.) A.J. Paton (both
members of subgen. Nautochilus (Bremek.) A.J. Paton).
In the study by Paton & al. (2004), the latter two species
formed a monophyletic group that separated from the
rest of Ocimum. Endostemon obtusifolius was selected as
an outgroup species to represent the clade sister to the
Ocimum/Syncolostemon/Hemizygia clade in the study of
Paton & al. (2004).
Species representative of variation in Hemizygia and
Syncolostemon were selected for sequence analysis
based on the results of a preliminary phylogenetic analy-
sis using morphological data, which grouped the species
into nine clades (results not shown). Nineteen species
representative of these nine clades were sampled for
sequence analysis. Appendix 1 lists voucher information
of plant samples and GenBank/EMBL accession num-
bers for all DNA sequences.
DNA extraction, PCR amplification and se-
quencing. —
Total DNA was extracted from fresh
leaves dried in silica gel (17 species) and in two cases (S.
comptonii and S. flabellifolius) from herbarium speci-
mens using the 2X CTAB method of Doyle & Doyle
(1987). The trnL-trnF region (hereafter trnL-F) and the
rps16 intron were amplified by PCR and sequenced.
Primers, PCR and sequencing reactions were outlined
earlier by Paton & al. (2004). In a few species where
amplification proved problematic, the primer pairs “c/d”
and “e/f” (Taberlet & al., 1991) were used. Sequences
were edited in Sequence Navigator and assembled in
“Autoassembler” (Perkin Elmer Applied Biosystems,
Warrington, U.K.). All base positions were checked to
ascertain agreement in the complementary strands before
assembling. Sequence alignment was done by eye. All
informative indels were added to the data matrix as addi-
tional binary characters in addition to being coded as
missing data. When either the 5´ or 3´ ends of overlap-
ping indels started or stopped at the same point, the
indels were treated as one unordered multistate character
but when neither of their ends had the same point of ori-
gin they were treated as different (Simmons &
Ochoterena, 2000).
Morphological character selection, defini-
tion and coding. —
Morphological characters used in
this study were scored from herbarium material obtained
from: BOL, J, K, LMU, NH, NU and PRE. Data from
herbarium material were supplemented by field data
where available. From the outset all character variation
was considered potentially informative of relationships.
However those characters that were found to: (1) be
invariable, (2) show considerable infraspecific variation
or (3) vary continuously to the extent that discrete char-
acter states could not be delimited, were excluded from
the analysis (Poe & Wiens, 2000). Forty-eight characters
were identified and are presented in Appendix 2.
Thirteen were coded as binary and 35 as multistate char-
acters. Thirty-eight of the characters are associated with
the inflorescence and flowers while the other ten are
derived from the leaves. The morphological data matrix
is provided in Appendix 3. In the matrix, missing data,
which includes unavailable as well as inapplicable data,
represent less than 2% of the entries.
Cladistic analyses. —
A total of 7 different analy-
ses were carried out. These were: (1) morphological data
for 54 taxa, comprising Hemizygia, Syncolostemon,
three representatives of Ocimum and the outgroup
species (E. obtusifolius); (2) morphological data but for
a restricted taxon set of 23 species for which both mor-
phological and molecular data were available, including
13 species of Hemizygia, six of Syncolostemon, three of
Ocimum and the outgroup species; (3) trnL-F sequence
data with and without coded gap characters for 23
species; (4) rps16 sequence data with and without coded
gap characters for 23 species; (5) combined trnL-F and
rps16 sequence data with and without coded gap char-
acters for 23 species; (6) combined molecular and mor-
phological data with and without coded gap characters
for 23 species; and (7) combined molecular and mor-
phological data with and without coded gap characters
for the complete taxon set of Hemizygia, Syncolostemon,
three representatives of Ocimum and the outgroup
species. Since the number of taxa included in the second
analysis was reduced, five morphological characters
whose scores were now constant for all the included
species were excluded, thus reducing the number of
morphological characters used to 43 instead of 48. In the
last analysis molecular data were unavailable for 31 ter-
minal taxa and were therefore represented by “missing
data” symbols in the matrix. The total number of char-
acters used in the combined analysis both for the
restricted and complete taxon set was 1665, of which 48
were morphological and 1617 molecular.
Maximum parsimony (MP) analyses were imple-
mented for all data sets using PAUP* (version 4.0b10,
Swofford, 2001). Heuristic searches were conducted
with 100 random addition replicate searches of equally
weighted (EW) characters (Fitch, 1971), TBR branch
swapping with MULPARS, the steepest descent option
not in effect, holding ten trees in each replicate and all
character transformations treated as equally likely. All
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
943
trees generated were swapped to completion. Successive
approximation weighting (SW, Farris, 1969) was imple-
mented based on the rescaled consistency index (RC,
Farris, 1989) using the best-fit criterion and a base
weight of one. The shortest Fitch trees were used as the
basis for calculating the initial weights after which
reweighting (using the reweighting option in PAUP*)
was carried out until the length of the tree did not change
in two successive rounds. Bootstrap (BS) support values
(Felsentein, 1985) were obtained from 1000 replicates of
sampling characters with equal probabilities using TBR
swapping but holding only five trees per bootstrap repli-
cate. Simple addition sequence was used to build trees,
with groups having frequencies greater than 50% being
retained in the final bootstrap consensus tree. We applied
the following arbitrary scale of bootstrap support, as used
by Van der Bank & al. (2002): weak 50–74%, moderate
75–84% and strong 85–100%. Measures of character fit
in the analyses were estimated using the consistency
index (CI) and retention index (RI, Farris, 1989). The
partition homogeneity test (Farris & al., 1995) with 100
replicates was run to assess the degree of incongruence
between data sets. Since the partition homogeneity test of
the molecular and morphological data was done for the
smaller subset of taxa, it was not repeated for the expand-
ed taxon set.
Historical biogeography. —
The phytogeo-
graphical regions of mainland Africa as depicted in Van
Wyk & Smith (2001) [based mainly on White (1983) and
subsequent modification by Clarke in Huxley & al.
(1998)] were used to study the geographic origin of
Hemizygia and Synclostemon. These included the
Afromontane and Afroalpine archipelago-like Regional
Centre of Endemism (AA), the Maputaland-Pondoland
Regional Mosaic (MP), the Kalahari-Highveld Regional
Transition zone (KH), the Zambezian Regional Centre of
Endemism (ZA), the Swahilian Regional Centre of
Endemism (SW), the Karoo-Namib Regional Centre of
Endemism (KN), the Guinea-Congolian/Sudanian Re-
gional Transition zone (GS), the Sudanian Regional
Centre of Endemism (SU) and the Sahel Regional Centre
of Endemism. India (I) and Madagascar (M) were treat-
ed, in a broad sense, as the geographic distributional
areas for H. comosa and H. madagascariensis respec-
tively. The Afromontane and Afroalpine archipelago-like
Regional Centre of Endemism in South Africa can fur-
ther be divided, following Van Wyk & Smith (2001) into
the Drakensburg Alpine, Barberton, Wolkberg, Soutspan-
berg and Chimanimani-Nyanga Centres.
Each geographic area was treated as a binary charac-
ter with two states, present or absent. Using either the
forward or reverse Camin-Sokal parsimony, these char-
acters were optimized using ACCTRAN option onto the
taxon cladogram to construct a taxon-area cladogram
[implemented in PAUP* version 4.0b10 (Swofford,
2001) using the irreversible character type option].
Bremer’s (1992) ancestral area analysis method was then
used to approximate ancestral areas of individual groups
from the topological information in the area cladogram
constructed. This was done by computing the presence or
gains (G) and absence or losses (L) for each area and
then estimating the G/L ratios for the same areas. The
area(s) with the highest G/L ratio was then considered to
be part of the hypothetical ancestral area.
RESULTS
Morphological data (complete and restrict-
ed taxon set). —
Maximum parsimony (MP) analysis
of the 48 morphological characters for the complete
taxon set of Hemizygia and Syncolostemon with equal
weighting (EW) yielded 140 trees (CI = 0.36, RI = 0.70).
The strict consensus tree from this analysis was com-
pletely unresolved and is not shown. Successive weight-
ing gave 175 trees (CI = 0.61, RI = 0.85) and the strict
consensus of these trees is presented in Fig. 1. Hemizygia
and Syncolostemon together form a monophyletic group
with moderate support (BS 76%). The strongly support-
ed (BS 88%) clade comprising subspecies of H. pretori-
ae, the weakly supported (BS <50%) clade in Fig. 1 of S.
flabellifolius with H. madagascariensis and H. sp. nov.
are unresolved at the base of the tree. H. albiflora is sis-
ter to the remainder of Hemizygia and Syncolostemon
(clade labeled I, BS <50%).
A strict consensus of trees from the EW MP analysis
of morphological data for the restricted taxon set was
largely unresolved with Hemizygia and Syncolostemon
forming a weakly supported (BS 50%) monophyletic
group (tree not shown). When the data was successively
weighted (SW), the topology of the strict consensus tree
is largely congruent with that obtained in the SW analy-
sis of the expanded data set (Fig. 1).
Separate and combined molecular data
(restricted taxon set). —
Alignment of the entire
trnL-F and rps16 regions resulted in 11 potentially phy-
logenetically informative indels (insertions/deletions):
four in the trnL-F region and seven in the rps16 region.
The indels ranged from 1-bp to 9-bp long and all except
four were non-overlapping. In the cases where overlap-
ping occurred, two 4-bp long indels overlapped but with
neither the 5´ nor 3´ ends having the same point of origin
and were therefore treated as different (Simmons &
Ochoterena, 2000). The other case of overlapping indels
involved a 5-bp with a 1-bp long indel, both of which had
their 3´ ends stopping at the same point and were treated
as one unordered multistate character (Simmons &
Ochoterena, 2000).
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
944
Analyses of the trnL-F data with and without the
indels added as presence/absence gap characters gave
similar results and this was also true for the rps16 data.
No hard incongruencies (Seelanan & al., 1997; Wiens,
1998) were detected in the results of the separate analy-
ses (data not shown), which, topologically, were also
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
945
Fig. 1. Strict consensus of 175 trees found after successively weighting morphological data of a complete taxon set of
Hemizygia and Syncolostemon. Bootstrap support > 50% is indicated above the branches (SW) and below in bold (EW).
88
82
69
78
67
62
57
60
66
76
64
74
67
72
52
77
62
50
93
86
69
56
86
60
E. obtusifolius
O. serrat um
O. labiatum
O. tenuiflorum
H. albiflora
H. bracteosa 1
H. bracteosa 2
H. bracteosa 3
H. welwitschii
H. canescens
H. petrensis
H. linearis
H. petiolata
H. elliottii
H. floccosa
H. persimilis
H. comosa
H. obermeyerae
H. parvifolia
H. cinerea
H. rehmannii
H. subvelutina
H. teucriifolia
H. oritrephes
H. incana
H. stenophylla
H. rugosifolia
H. gerrardii
H. thorncroftii
H. foliosa
H. transvaalensis
S. ramulosus
S. argenteus
S. eriocephalus
S. concinnus
S. parviflorus
S. sp. nov.
S. comptonii
S. densiflorus
S. rotundifolius
H. macrophylla
S. latidens
S. macranthus
H. stalmansii
H. ornata
H. ramosa
H. modesta
H. punctata
H. bolusii
H. pretoriae 1
H. pretoriae 2
S. flabellifolius
H. madagascariensis
H. sp. nov.
80
similar. The two data sets were hence combined. The
combined data matrix without coded gap characters had
1606 variable sites, 23 of which were parsimony inform-
ative (Table 1). Equal weights analysis of this data pro-
duced 6 trees (CI = 0.96, RI = 0.95) and with SW also
gave six trees (CI = 1.0, RI = 1.0). The results of both
analyses were congruent (Fig. 2). When analyzed with
EW and SW but now including coded gap characters, the
data gave similar results.
Combined molecular and morphological
data (restricted taxon set). —
The partition homo-
geneity test of the individual molecular and morphologi-
cal data sets for the restricted taxon set of 23 species for
which both data sets were available, gave a P value =
1.000. This suggests that the combined molecular data
set was highly congruent with the morphological one.
The only incongruence in the results from the separate
EW analyses of the two data sets was in the position of
H. subvelutina (data not shown). The incongruence was,
however, not of the “hard” type (Seelanan & al., 1997;
Wiens, 1998) meaning that incongruent topologies were
not supported by bootstrap values of over 50%.
Similarly, in the SW analyses, there was incongruence in
the positions of H. punctata, H. pretoriae, H. obermey-
erae, H. subvelutina and S. macranthus but these too
were not of the “hard” type. Thus the two data sets were
combined since none of the incongruencies was strongly
supported by both data sets (Seelanan & al., 1997;
Wiens, 1998; Reeves & al., 2001).
Equal weights analysis of the combined data with
coded gap characters included yielded 33 trees (CI =
0.66, RI = 0.68) while SW analysis produced five trees
(CI = 0.90, RI = 0.90). Analysis of the same data with
EW but without gap characters produced 15 trees (CI =
0.66, RI = 0.68). With the characters successively
weighted two trees (Table 1; CI = 0.90, RI = 0.89) were
obtained. Results from the SW MP analyses of the com-
bined data with coded gap characters and without them
(data not shown) were basically congruent.
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
946
Table 1. Summary of tree statistics for the individual and combined data sets. Superscript “a” indicates data of a
restricted taxon set, “b” indicates data of a complete taxon set, “c” indicates data excludes gaps coded as pres-
ence/absence characters and “d” indicates data including gaps coded as presence/absence characters.
No. of Length of No. of No. of
informative shortest trees resolved CI CI RI RI
Data analysed characters tree (EW) (SW) nodes (SW) (EW) (SW) (EW) (SW)
trnL-F sequencesac 10 44 5 5 0.97 1.0 0.95 1.0
trnL-F sequencesad 13 47 10 5 0.97 1.0 0.96 1.0
rps16 sequencesac 13 56 5 7 0.87 1.0 0.89 1.0
rps16 sequencesad 20 57 5 7 0.88 0.93 0.89 0.94
trnL-F + rps16 sequencesac 23 85 6 8 0.96 1.0 0.95 1.0
trnL-F + rps16 sequencesad 33 101 6 8 0.92 1.0 0.90 1.0
morphological dataa37 114 6 13 0.52 0.77 0.66 0.85
morphological datab41 191 175 29 0.36 0.61 0.70 0.85
combined mol. + morph. dataac 71 214 2 18 0.66 0.90 0.68 0.89
combined mol. + morph. dataad 61 233 33 16 0.66 0.90 0.68 0.90
combined mol. + morph. databc 74 287 55 33 0.52 0.83 0.71 0.87
combined mol. + morph. databd 64 309 130 28 0.52 0.83 0.71 0.88
Fig. 2. Strict consensus of six most parsimonious trees
found after successively weighting combined trnL-F and
rps16 sequence data of a restricted taxon set of
Hemizygia and Syncolostemon with presence/absence
gap characters excluded. Bootstrap support >50% is indi-
cated above branches (SW) and below in bold (EW).
Dotted lines represent branches not supported by EW
analysis of data when gap characters are included.
71
63
61
92
66
64
64
61
86
98
95
E. obtusifolius
O. labiatum
O. serratum
O. tenuiflorum
H. albiflora
H. incana
H. parvifolia
H. punctata
S. macranthus
H. obermeyerae
H. teucriifolia
H. subvelutina
H. transvaalensis
H. pretoriae
H. foliosa
H. modesta
H. stalmansi
i
S. argenteus
S. parviflorus
S. comptoni
i
S. rotundifolius
H. persimilis
S. flabellifolius
63
98
95
86
64
65
82
63
Combined molecular and morphological
data (complete taxon set). —
Maximum parsimony
analysis of the combined data for the complete taxon set
with EW but minus coded gap characters produced 25
trees (CI = 0.52, RI = 0.71) while SW analysis gave 55
trees (Table 1, CI = 0.83, RI = 0.87). The strict consen-
sus trees from both the EW and SW analyses were iden-
tical (Fig. 3). In both, Hemizygia and Syncolostemon
were retrieved together as a monophyletic group (BS
71%, EW; 87%, SW). Generally, BS support values of
internal nodes in the SW analysis were higher than in the
EW analysis (Fig. 3). When the gap characters were
included and the data analyzed with EW, five trees (CI =
0.52, RI = 0.71) were produced. The strict consensus tree
was identical to the one produced when the data were
successively weighted and analyzed without the coded
gap characters. On reweighting, 130 trees (Table 1; CI =
0.83, RI = 0.88) were produced while the strict consen-
sus tree (not shown) was different from the one obtained
in the EW analysis. There were no hard incongruencies
(Seelanan & al., 1997; Wiens, 1998).
Morphological characters on the phylogeny.
The CI and RI of morphological characters on the MP
EW tree shown in Fig. 3 are presented in Appendix 4. Of
the 48 morphological characters, only seven are not
homoplasious. Fused anterior stamens (Fig. 4I) are syn-
apomorphic for the Hemizygia/Syncolostemon clade
while the S. macranthus-S. ramulosus clade is diagnosed
by two synapomorphies (Table 2): a free posterior calyx
lip (Character 20, Fig. 4A) which is not raised at all or if
raised, very weakly so (Character 22, Fig. 4A). Brown-
ish-black to black nutlets (Character 37) are synapomor-
phic for the H. canescens-H. petiolata subclade while the
H. elliottii-H. comosa subclade is diagnosed by filaments
of the anterior stamens that are fused only below the mid-
dle (Character 32). Homoplasious characters diagnose all
other clades and subclades denoted on Fig. 3. Some
clades share certain diagnostic characters (Table 2) but
still differ in others. Characters like the presence of a
style shield, truncate apex of nutlet, hairs on the nutlet
apex and an obtriangular nutlet cross-section are autapo-
morphies in the outgroup. Mature nutlets 0.5–1 mm long
and the presence of equal lobes on the disk are also
autapomorphies in Ocimum tenuiflorum (Appendix 4).
Historical biogeography. —
Based on the results
of the combined data analysis with EW (Fig. 3) and ap-
plying Bremer’s (1992) method, the Afromontane and
Afroalpine archipelago-like Centre of Endemism showed
the highest gain to loss ratio of 1.05 (Table 3), followed
by Madagascar (M, 1.00), the Zambezian Regional Cen-
tre of Endemism (ZA, 0.79) and the Kalahari-Highveld
Regional Transition zone (KH, 0.61). Of the outgroup
species E. obtusifolius, O. serratus and O. labiatus are
widespread in the Kalahari-Highveld Regional Transition
zone, the Zambezian Regional Centre of Endemism and
the Maputaland-Pondoland Regional Mosaic (MP) while
O. tenuiflorum is widespread in Asia. If the ancestral area
analysis is rerun to include the distribution of the out-
groups, the value for Madagascar drops to 0.25, while the
values for other areas do not change significantly.
If the Afromontane and Afroalpine archipelago-like
centre of endemism is further divided and G/L ratios cal-
culated for the delimited areas, then the Barberton and
adjacent Wolkberg Centres have the highest G/L values
(Table 3).
DISCUSSION
Relationship between Hemizygia and Synco-
lostemon. —
Monophyly of a clade that encompasses
both Hemizygia and Syncolostemon is supported by all
analyses except the EW analysis of morphological data
using the complete taxon set, which was unresolved. The
group is united by a single morphological synapomor-
phy, i.e. the presence of fused anterior stamens (Fig. 4I).
Neither Hemizygia nor Syncolostemon as currently
defined is therefore monophyletic. Only one mono-
phyletic genus is thus recognized here using the earlier
name Syncolostemon. New combinations following this
treatment are given at the end of this paper.
Separate versus combined analysis. —
Most
clades common to the morphological and molecular trees
and also present in the combined data tree (Fig. 3)
showed an increase in BS values in the latter. These areas
of agreement are therefore likely to represent real groups
(de Queiroz & al., 1995). That there was largely no
decrease in BS values is an indication that the data sets
did not have any incongruence driven by biological phe-
nomena (Adreasen & Bremer, 2000). The small number
of informative molecular characters (Table 1) appears
not to have been swamped by the morphological data.
Instead, in different parts of the combined data tree (Fig.
3), they seem to complement the morphological data by
either increasing clade support and/or clarifying phylo-
genetic relationships. The results from the combined
analysis therefore seem to support the view that when
data are combined, more robustly supported and resolved
topologies can be obtained (e.g., Kluge, 1989; Donoghue
& Sanderson, 1992; Soltis & al., 1998; Andreasen &
Bremer, 2000; Eldenäs & Linder, 2000; Klompen & al.,
2000; Reeves & al., 2001; Goldblatt & al., 2002).
To a large extent, the EW and SW analyses of the
combined data with coded gap characters excluded using
the complete taxon set (Fig. 3), recovered more or less
the same clades and same relationships within clades as
the SW analysis of morphological data (Fig. 1), and the
differences did not reflect hard incongruencies. Given
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
947
that the diagnoses of these clades are not ambiguous
since the characters defining them are not again reversed,
they are likely, on the basis of total evidence and within
the context of strict monophyly (Buckland & Bremer,
1998; Carine & Scotland, 2002), to be potential lineages
within Syncolostemon s.l. However since most of them
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
948
Fig. 3. One of the 55 most parsimonious trees of combined molecular and morphological data of a complete taxon set
of Hemizygia and Syncolostemon found after successive weighting with presence/absence gap characters excluded.
Geographical distribution areas follow names of taxa. Refer to Table 3 for acronyms. Bootstrap support values >50%
are indicated above branches (SW) and below in bold (EW). Dotted lines represent branches collapsing in the strict
consensus analysis of both SW and EW trees. Letters refer to clades that are strongly supported and discussed in the
text.
72
87
71
87
68
88
53
83
67
86
54
74
57
77
54
51
65
96
55
94
56
51
94
63
89
60
60
82
87
89
79
E. obtusifolius
O. labiatum
O. serratum
O. tenuiflorum
H. albiflora KH ZA
H. rehmannii ZA
H. incana KH ZA
H. parvifolia KH ZA
H. cinerea AA
H. obermeyerae ZA
H. rugosifolia ZA
H. subvelutina KH ZA
H. teucriifolia AA ZA
H. oritrephes AA
H. stenophylla AA MP
H. gerrardii AA
H. bracteosa 1 KH MP SW SA ZA
H. bracteosa 2 GS SA SU
H. bracteosa 3 ZA
H. welwitschii GS ZA
H. canescens AA KH ZA
H. petrensis KH ZA
H. linearis KH ZA
H. petiolata AA KH ZA
H. transvaalensis AA KH ZA
H. pretoriae 1 KH ZA
H. pretoriae 2 AA
H. ramosa MP
H. foliosa AA KH
H. thorncroftii AA ZA
H. ornata ZA
H. stalmansii AA KH
S. ramulosus MP
S. argenteus AA MP
S. eriocephalus KH ZA
S. parviflorus AA KH MP
S. concinnus AA KH ZA
S. comptonii MP ZA
S. sp. nov. AA MP
S. densiflorus AA MP
S. rotundifolius AA MP
H. macrophylla KH
S. latidens AA
S. macranthus AA
H. modesta KH ZA
H.
p
unctata AA KH ZA
H. bolusii AA
H. elliottii AA ZA
H. floccosa KN
H.
ersimilis AA KH
H. comosa I
H. s
. nov. SW ZA
S. flabellifolius AA
H. madagascariensis M
D
B
A
C
89
93
lack support, we do not formally recognize them here.
Further discussion is therefore restricted to groups that
are strongly supported by our data (Fig. 3).
Clade A. —
The results strongly support (BS 94%)
Hemizygia subvelutina and H. teucriifolia as sister taxa
and these findings thus seem to confirm Codd’s (1976b,
1985) observation that the two species are closely relat-
ed. In both the corolla tubes are slightly constricted at the
mouth (Fig. 4H), the posterior and anterior stamens
rarely exceed the length of the lower corolla lip (charac-
ters 29 and 31) and the cymes are subtended by persist-
ent bracts (character 14). However, on herbarium speci-
mens and even in the field, the distinctive yellow colour
of the tomentum on the leaves and branches of H. subve-
lutina decisively separate them. In addition, the leaves of
H. subvelutina are narrow with very strongly revolute
margins though in some odd specimens they may also be
broad. In H. teucriifolia the leaves are generally broad.
Clade B. —
Even though the H. canescens-H. peti-
olata lineage receives strong support (BS 94%), we do
not formally recognize it as this would require the recog-
nition of other poorly supported clades in the tree as well.
Nonetheless the results suggest a close relationship
between species in this clade. It is diagnosed by the pres-
ence of black nutlets (a unique synapomorphy for the
clade), absence of sessile glands on the leaves and pos-
session of a slightly upcurved corolla tube which is par-
allel-sided from the base to near the mouth from where it
dilates (Fig. 4G). Relationships between H. canescens,
H. petrensis and H. linearis are unresolved and the three
emerge in a weakly supported (BS 51%) subclade. They
all occur in more or less similar habitat and altitudinal
ranges (Codd, 1976b, 1985). Hemizygia petiolata has
many characteristics in common with the complex of
three species discussed above and is also strongly sup-
ported (BS 93%) as sister to this complex by morpho-
logical data (see Fig. 1). However, it differs in having a
strong smell of mint and coconut (Codd, 1985), bracts
that sometimes persist as a coma, leaves that are strong-
ly ovate with comparatively longer petioles and longer
corolla tubes.
Clade C. —
This clade represents the two sub-
species of Hemizygia pretoriae (subsp. pretoriae, subsp.
heterotricha) which normally, are differentiated on the
basis of the latter having branched instead of simple hairs
(Codd, 1976b, 1985). However in some specimens of the
latter, e.g., Schrire 1170 (NH), simple hairs are more
abundant on the stems, leaves and bracts than dendroid
hairs which raises questions concerning the delimitation
of these taxa by Codd (1976b, 1985). Hemizygia pretori-
ae resembles H. persimilis in habit and ecology (Codd,
1976b, 1985) and some aspects of morphology, e.g., the
persistent bracts subtending the cymes and included pos-
terior stamens. However none of our analyses suggests a
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
949
Table 2. Diagnostic characters for clades in the most parsimonious tree shown in Fig. 4. Characters marked with * are unique synapomorphies.
H. albiflora- H. gerrardi H. bracteosa1- H. transvaa- H. pretoriae1- H. foliosa- H. stalmansii S. ramulosus- H. modesta- H. elliottii-
Clade H. stenophylla H. petiolata lensis H. ramosa H. ornata S. macranthus H. bolusii H. sp. nov.
Leaf attachment - - - - - sessile - - - sessile
Leaf margin revolute - - - - - revolute - - -
Disposition of sessile glands - - - - sunken - - sunken - -
Branching of inflorescence - - - - - - simple but with - - -
1 or 2 pairs of
basal branches
Terminal bracts of inflorescence - - - - - - - absent - present
Throat of calyx - - - - - ? setose - - -
Attachment of posterior calyx lip - - - - - ? - free* - -
Orientation of posterior calyx lip - - - - - ? - not raised* -
Corolla tube expansion - parallel-sided from - - - ? - - - -
base up to near mouth
where it dilates
Insertion of posterior stamen - above the middle - - above the middle ? - - - -
of corolla of corolla
Apex of stylopodium - - lobed - - - - - - -
Style remains atapex of stylopodium - - sunken - - - - - - -
Thin membranes hanging from - - - - - - - absent - -
stylopodium
close relationship between them. It is therefore possible
that the characters they have in common may have aris-
en independently in each, through parallel evolution.
Clade D. —
The monophyly of this clade is very
strongly supported (BS 89%). Two unique synapomor-
phies and three homoplasious characters, one of which is
rare in Syncolostemon s.l., diagnose the clade (see Table
2). Though also seen in S. flabellifolius, the naked sty-
lopodium (Fig. 4E) is an apomorphy in the clade.
Hitherto, members of this clade, with the exception of H.
macrophylla, have been treated as Syncolostemon s.s.
The clade can further be partitioned into subclades.
Syncolostemon latidens and S. macranthus form a weak-
ly supported subclade (BS 65%), which has strong sup-
port (BS 89%) as sister to the remainder of the clade.
This subclade is diagnosed by an elliptic to broadly ellip-
tic posterior lobe of the calyx (Fig. 4C). On the basis of
the shape of the posterior lobe of the calyx, S. macran-
thus and S. latidens occupy an intermediate position
between Hemizygia and Syncolostemon sensu Codd
(1976a). Both species are rather restricted in distribution
but occur in different areas within KwaZulu-Natal.
Syncolostemon macranthus is largely confined to the
Drakensberg area while S. latidens is only known from a
few collections around Kranskop.
Hemizygia macrophylla originally described as
Syncolostemon macrophyllus by Gürke (1898 cited in
Codd, 1976a), was transferred to Hemizygia on the basis
of its broadly ovate upper lip of the calyx (Codd, 1976a)
though no mention was made of whether this was decur-
rent on the calyx tube or not. Generally, the mature calyx
in this species is setose in the throat and constricted at the
mouth (Codd, 1976b). The latter character is diagnostic
of the clade comprising H. macrophylla and the remain-
der of the S. macranthus-S. ramulosus clade. From all
other species in Hemizygia, H. macrophylla is distin-
guished by the setose calyx throat and the lax, branched
inflorescence (Codd, 1976b). Significantly, our examina-
tion of this species revealed the upper lip of the calyx to
be non-decurrent and not raised as in other species with-
in the clade. It is considered to have a distant relationship
with S. parviflorus (Codd, 1976a) and this is confirmed
by its moderately supported (BS 82%) position as sister
to the remainder of the clade. Thus, contrary to Codd
(1976b), H. macrophylla, is clearly more closely related
to members of this clade than to other species placed in
Hemizygia by Codd. From species of Syncolostemon in
the clade, it is distinguished by the velvety covering of
branched hairs on the leaves and the obovate to ovate
upper lip of the calyx.
The S. rotundifolius-S. ramulosus clade is a moder-
ately supported (BS 82%) clade diagnosed by a single
synapomorphy, i.e., presence of a deltoid to deltoid-sub-
ulate posterior lip of the calyx (Fig. 4B). All species in
this clade also lack glandular hairs. However,
Syncolostemon rotundifolius is distinguished from the
rest of the clade by its 2-flowered verticils.
Syncolostemon ramulosus, which forms a weakly
supported (BS 60%) clade with the remainder of the S.
macranthus-S. ramulosus clade, has previously been
synonymized under S. densiflorus (Codd, 1976a, 1985).
However, Codd (1988) reinstated it to specific rank. It
forms part of the group Codd (1976a) designated as
Group 2 of Syncolostemon. The presence of leaf fascicles
on short shoots arising from the axils of other leaves in
S. ramulosus distinguishes it from S. rotundifolius with
which it has also been much confused (Codd, 1985).
Syncolostemon densiflorus is weakly supported (BS
60%) as sister to a strongly supported (BS 89%) but
unresolved clade of six species. It is distinguished from
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
950
Table 3. Ancestral area analysis following Bremer (1992). The most probable ancestral area(s) (AA) correspond to the
highest G/L quotient. For AA subdivisions and species are indicated.
Area Gains Losses Gains/Losses
Afromontane and Afroalpine (AA) 22 21 1.05
Wolkberg Centre (H. albiflora, H. bracteosa1, H. canescens, H. parvifolia, H. persimilis, H. rehma- 7 18 0.38
nii, H. stalmansii, H. subvelutina, H. thorncroftii, H. transvaalensis, H. petiolata, H. rugosifolia)
Baberton Centre (H. albiflora, H. bracteosa1, H. canescens, H. incana,H. modesta, H. parvifolia, 10 28 0.36
H. persimilis, H. pretoriae, H. punctata, H. stalmansii, H. thorncroftii, H. transvaalensis, H. teucri-
folia, S. concinnus, H. comptonii, S. parviflora
Soutspansberg Centre (H. bracteosa1, H. petiolata, H. petrensis, H. elliotii, H. canescens, H. petiolata) 3 18 0.16
Chimanimani-Nyanga Centre (H. oritrepes, S. flabelliformis) 2 12 0.16
Drakensberg Alpine Centre (H. cinerea, S. macranthus, H. bolusii) 2 17 0.12
Madagascar (M) 1 1 1.00
Zambezian (ZA) 19 24 0.79
Kalahari-Highveld (KH) 16 26 0.61
Maputaland-Pondoland (MP) 7 15 0.47
India (I) 1 5 0.20
Sahel (SA) 2 10 0.20
Swahilian (SW) 2 12 0.16
Karoo-Namib (KN) 1 7 0.14
Sudanian (SU) 1 11 0.09
the members of this unresolved clade by the presence of
petiolate leaves, a tubular calyx and the absence of the
frill-like structure at the base of the nutlets. The unre-
solved clade of six species is diagnosed by a corolla that
dilates from the base (Fig. 4F) but this character is homo-
plasious (see Appendix 4) and hence of limited diagnos-
tic value (Carine & Scotland, 2002). Codd (1976a) des-
ignated the six-taxon polytomy (excluding S. sp. nov.) as
Group 1 of Syncolostemon. It has also previously been
recognized as section Micranthi (Briquet, 1897). Even
though relationships between members of the six-taxon
polytomy are not resolved (Fig. 3), in the analysis of
combined data using a restricted taxon set (data not
shown), three species in the group viz. S. argenteus, S.
parviflorus and S. comptonii form a very strongly sup-
ported (BS 100%) monophyletic group. This suggests the
existence of very close affinities between members of the
group. However the lack of resolution in the combined
data analysis using the complete taxon set (Fig. 3) is
probably because only three out of the six species had
molecular data.
It is not always easy to differentiate between some
specimens of Syncolostemon parviflorus and S. argen-
teus from the Pietermaritzburg area where the two
species overlap and where it is also suspected they may
be undergoing introgression (Codd, 1976a). Both species
occur in dense grassland and often have many stems
growing from a woody rootstock, a characteristic associ-
ated with frequent grassland fires. However, the consis-
tently appressed hairs on the leaves of S. argenteus are
decisive in separating them. Syncolostemon comptonii
also closely resembles S. parviflorus, but its compact,
branched inflorescence and robust habit clearly distin-
guishes it. Hitherto it is only known from a few speci-
mens collected in Swaziland.
Syncolostemon eriocephalus Verdoorn forms a
weakly supported (BS 63%) clade with S. argenteus.
Sericeous leaves with appressed simple hairs and a vil-
lous throat of the calyx, both of which are rare in Syn-
colostemon, diagnose the clade. However the shrubby
habit, inflorescence borne terminally on short lateral
shoots and the calyx, which is completely covered with
yellowish-white hairs, distinguish S. eriocephalus from
S. argenteus. Also, S. eriocephalus occurs more norther-
ly, in the Limpopo and Mpumalanga Provinces, while S.
argenteus is found more southerly, in KwaZulu-Natal,
occurring in shallow sandy soils among quartzite rocks
(Codd, 1976a; 1985). Syncolostemon concinnus can be
easily confused with S. parviflorus in facies. However,
branched hairs on the leaves distinguish it from all the
other species in the six-taxon clade.
Syncolostemon sp. nov., an entity from KwaZulu-
Natal, has mistakenly been identified, on the only speci-
men available, as S. parviflorus despite being strikingly
different in corolla morphology. The corolla tube is long
and parallel-sided from the base and only dilates close to
the mouth (character 24) unlike in S. parviflorus where it
dilates straight from the base towards the mouth (e.g.,
Fig. 4F). Also, it has a setose interior (character 25)
instead of a ring of hairs as in S. parviflorus. The mor-
phology of the corolla tube in this entity is suggestive of
an adaptation to a different pollinator.
Biogeography. —
Results from the ancestral area
analysis (Bremer, 1992) give the Afroalpine and
Afromontane archipelago-like Regional Centre of
Endemism the highest G/L ratio (Table 3) suggesting it
may be the ancestral area for Syncolostemon s.l. The high
G/L ratio for Madagascar only demonstrates that one of
the basal sister groups of Syncolostemon s.l. is in
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
951
Fig. 4. Calyx, floral and disk morphology. A, calyx sho-
wing lobes of more or less equal size and weakly raised,
free upper lip (fr) (S. concinnus); B, spread out calyx sho-
wing deltoid upper lip (d) and deltoid-subulate lateral and
lower lobes (S. densiflorus); C, spread out calyx showing
non-decurrent, ovate-elliptic upper lip and deltoid lateral
and lower lobes (S. macranthus); D, calyx showing stron-
gly raised and decurrent upper lip (dct) and subulate to
spinescent lobes of lower lip (H. bracteosa); E, disk sho-
wing naked stylopodium (S. concinnus). F, corolla tube
diverging from the base to a wide mouth (S. rotundifoli-
us). G, upcurved corolla tube, parallel-sided from the
base to near the mouth where it diverges (H. petiolata). H,
corolla dilating from the base but with a distinct con-
striction at the mouth (H. subvelutina). I, two anterior sta-
mens showing position of fusion (sf) (S. rotundifolius).
Scale bars: A–H = 1 mm, I = 0.5 mm.
Madagascar while the other is not (Fig. 3). It is, howev-
er, much reduced when the distribution of the outgroups
is considered. A broader biogeographic analysis across
the tribe by Paton & al (2004) suggests the ancestral area
may be in Africa but this analysis did not include
Hemizygia madagascariensis. Views that Madagascar
separated from Africa during the mid-Cretaceous (Leroy,
1978) and that the first elements of tribe Ocimeae arrived
in Africa during the Paleogene (Raven & Axelrod, 1974)
do not also favour a hypothesis of Madagascar as the
ancestral area since the arrival of H. madagascariensis in
Madagascar can only be presumed to have occurred after
Madagascar separated from Africa.
Syncolostemon s.l. is not found throughout the Afro-
alpine and Afromontane archipelago-like Regional Cen-
tre of Endemism. This area extends from the southern
Drakensberg of South Africa, through the high moun-
tains of Eastern Africa and north to the Simian
Mountains of Ethiopia. Within this large region
Syncolostemon s.l. is found only as far north as the
Chimanimani-Nyanga Centre in Zimababwe (Van Wyk
& Smith, 2001). Examining the distribution of the
species found within the Afroalpine and Afromontane
archipelago-like Regional Centre of Endemism suggests
that the adjacent Wolkberg and Barberton Centres
include the most species of Syncolostemon s.l. An AA
analysis of the mountain divisions of the Afroalpine and
Afromontane archipelago-like regional centre of end-
emism in southern Africa suggests that these areas are
most likely to comprise the ancestral area of
Syncolostemon s.l. (Table 3).
Given that all other members of Syncolostemon s.l.,
with the exception of Hemizygia madagascariensis, are
African, it is rather puzzling how H. comosa, which only
occurs in India, may have attained its present distribu-
tion. Recent analyses have depicted India as being close
to Africa even by the upper Jurassic [about 140 Myr]
(Weijermars, 1989). Its journey away from Africa may
have begun in the late Jurassic [about 148 Myr]
(Williams, 1986), though others (e.g., Rabinowitz & al.,
1983) suggest it may have been even earlier, about 165
Myr. By the early Cretaceous, India was relatively iso-
lated from Africa and had started moving northward on a
course close to the African continent (Weijermars, 1989).
It is depicted in the Palaeocene (about 60 Myr), as
approaching Eurasia but remaining close to or in contact
with Africa (Briggs, 1989). The Owen Fracture Zone, a
north-south ridge that runs close to the tip of the Somali
Peninsula, probably marks the track of India’s western
margin as it moved northward (Briggs, 1987). Elements
of tribe Ocimeae which are considered to have first
arrived in Africa in the Paleogene (Raven & Axelrod,
1974) may have dispersed to India during this period
when India is postulated to have been in close contact
with Africa (Briggs, 1987, 1989). It is possible that H.
comosa represents a very recent arrival of Hemizygia on
the sub-continent as has also been suggested for other
members of the Ocimeae like Platostoma, Orthosiphon,
Plectranthus and Anisochilus (Paton & al., 2004). From
the foregoing it is evident that dispersal events may have
played a significant role in dictating the present disjunct
distribution of Syncolostemon s.l.
TAXONOMY
Syncolostemon E. Mey. ex Benth. in E. Mey., Comm.
230 (1837). – Lectotype (Codd, 1976b): S. parvi-
florus E. Mey. ex Benth. (= Ocimum sect. Hemizygia
Benth., DC. Prod. 12: 41 (1848) = Hemizygia
(Benth.) Briq. Pflanzenfam. 4, 3a: 368 (1897) = Or-
thosiphon sensu Baker, Fl. Trop. Afr. 5: 365 (1900)
= Bouetia A. Chev., Mem. Soc. Bot. Fr. 2: 200
(1917).
New combinations in Syncolostemon:
Syncolostemon albiflorus (N.E. Br.) D.F. Otieno, comb.
nov. Orthosiphon albiflorus N.E. Br. in Fl. Cap. 5,
1: 251 (1910) Hemizygia albiflora (N.E. Br.) M.
Ashby. J. Bot., Lond. 73: 348 (1935); Codd in
Bothalia 12: 10 (1976b); Codd in Fl. southern Afr.
30, 4: 201 (1985). Type: Transvaal, Pilgrims Rest
District, Mac Mac, Mudd s.n. (holotype, K!).
=O. decipiens N.E. Br. Fl. Cap., 5, 1: 252 (1910).
Type: Transvaal, Mac Mac, Mudd s.n. (holotype,
K!).
Syncolostemon bolusii (N.E. Br.) D.F. Otieno, comb.
nov. Orthosiphon bolusii N.E. Br. in Fl. Cap. 5, 1:
258 (1910) Hemizygia bolusii (N.E. Br.) Codd in
Bothalia 8: 159 (1964); Codd in Bothalia 14: 13
(1976b); Codd in Fl. southern Afr. 30, 4: 204 (1985).
Type: Natal, Giants Castle 6 Nov. 1897, A. Bolus in
Herb. Guthrie 4894 (holotype, BOL!)
Syncolostemon comosus (Wright ex Benth.) D.F.
Otieno, comb. nov. Orthosiphon comosus Benth in
DC., Prodr. 12: 52 (1848) Hemizygia comosa
(Wright ex Benth.) A.J. Paton in Kew Bull. 53: 2
(1998). Type: India, Wright 2567 (holotype, K!).
Syncolostemon elliottii (Bak.) D.F. Otieno, comb. nov.
Orthosiphon elliottii Bak. in Fl. Trop. Afr. 5: 376
(1900) Hemizygia elliottii (Bak.) M. Ashby in J.
Bot. 73: 345 (1935), pro parte excl. Natal spec. Codd
in Bothalia 12: 7 (1976b); Codd in Fl. southern Afr.
30, 4: 198 (1985). – Type: Matabeleland, Elliott s.n.
(holotype, K!).
=O. messinensis Good in J. Bot. 63: 173 (1925). Type:
Transvaal, Messina, Moss & Rogers 153 (holotype,
BM; isotype, PRE).
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
952
Syncolostemon foliosus (S. Moore) D.F. Otieno, comb.
nov. Orthosiphon foliosus (S. Moore) N.E. Br. in
Fl. Cap. 5, 1: 243 (1910) Hemizygia foliosa S.
Moore in J. Bot. 43: 172 (1905); Codd in Bothalia
12: 15 (1976b); Codd in Fl. southern Afr. 30, 4: 206
(1985). – Type: Swaziland, Mbabane, Burtt Davy
2833 (holotype, BM; isotypes, K!; PRE!).
=O. humilis N.E. Br., Fl. Cap., 5, 1: 259 (1910)
Hemizygia humilis (N.E. Br.) M. Ashby in J. Bot. 73:
348 (1935). – Type: Transvaal, Waterval Onder,
Rogers 4375 (holotype K; isotype, PRE).
Syncolostemon floccosus (Launert) D.F. Otieno, comb.
nov. Hemizygia floccosa Launert in Mitt. Bot.
Munchen 7: 302 (1968); Codd in Bothalia 12: 8
(1976b); Codd in Fl. southern Afr. 30, 4: 198 (1985).
Type: S.W.A./Namibia, Outjo, De Winter & Hardy
8139 (holotype, PRE; isotype, M).
Syncolostemon gerrardii (N.E. Br.) D.F. Otieno, comb.
nov. Orthosiphon gerrardii N.E. Br. in Fl. Cap. 5,
1: 249 (1910) Hemizygia gerrardii (N.E. Br.) M.
Ashby in J. Bot. 73: 345 (1935); Codd in Bothalia
12: 8 (1976b); Codd in Fl. southern Afr. 30, 4: 198
(1985) Type: Natal. “near Ingoma”, Gerrard 1239
(holotype, K).
Syncolostemon incanus (Codd) D.F. Otieno, comb. nov.
Hemizygia incana Codd in Bothalia 12: 5 (1976b);
Codd in Fl. southern Afr. 30, 4: 197 (1985). – Type:
Transvaal, Kaapsche Hoop, Codd 5758 (holotype,
PRE!).
Syncolostemon madagascariensis (A.J. Paton & Hedge)
D.F. Otieno, comb. nov. Hemizygia madagas-
cariensis A.J. Paton & Hedge in Flore de
Madagascar et des’ Comores, 110 (1999). – Type:
Province de Fianarantsoa, Itremo, sur la RN 35, 50
km à l’ ouest d’ Ambatofinandrahana, sommet, 1670
m, Phillipson, Clement and Rafamantanantsoa 3866
(holotype, E; isotypes, K; MO, P, TAN).
Syncolostemon modestus (Codd) D.F. Otieno, comb.
nov. Hemizygia modesta Codd in Bothalia 12: 12
(1976b); Codd in Fl. southern Afr. 30, 4: 202 (1985).
Type. Swaziland, Mbabane, Bomvu Ridge, Compton
28368 (holotype, PRE!).
Syncolostemon obermeyerae (M. Ashby) D.F. Otieno,
comb. nov. Hemizygia obermeyerae M. Ashby in
J. Bot. 73: 343 (1935); Codd in Bothalia 12: 4
(1976b); Codd in Fl. southern Afr. 30, 4: 196 (1985).
– Type: Soutpansberg, Entabeni, Obermeyer sub
TRV 31556 (holotype, PRE!).
Syncolostemon oritrephes (Wild) D.F. Otieno, comb.
nov. Hemizygia oritrephes Wild in Kirkia 4: 149
(1964). – Type: Southern Rhodesia/Zimbabwe,
Chimanimani, 22 Sept. 1956, Taylor 1816 (holotype,
SRGH).
Syncolostemon parvifolius (Codd) D.F. Otieno, comb.
nov. Hemizygia parvifolia Codd in Bothalia 12: 4
(1976b); Codd in Fl. southern Afr. 30, 4: 197 (1985).
– Type: Transvaal, Farm Belvedere, overlooking
Blyde River Gorge, Codd 10321 (holotype, PRE!).
Syncolostemon persimilis (N.E. Br.) D.F. Otieno, comb.
nov. Orthosiphon persimilis N.E. Br. in Fl. Cap. 5,
1: 246 (1910) Hemizygia persimilis (N.E. Br.) M.
Ashby in J. Bot. 73: 349 (1935); Codd in Bothalia
12: 15 (1976b), Codd in Fl. southern Afr. 30, 4: 206
(1985). – Lectotype (Ashby 1936): Transvaal,
Barberton, Thorncroft sub TRV 3132 (lectotype K;
isolectotypes, PRE, SAM).
=O. rogersii N.E. Br., Fl. Cap., 5,1: 247 (1910).
Syntype: Transvaal, Nelspruit, Rogers sub TRV
4740 (syntypes K, PRE, SAM); Duiwels Kantoor,
Kaapschehoop, Bolus 9742 (syntype, BOL).
Syncolostemon pretoriae (Gürke) D.F. Otieno, comb.
nov. Orthosiphon pretoriae Gürke in Bot. Jahrb.
26: 81 (1898) Hemizygia pretoriae (Gürke) M.
Ashby in J. Bot. 73: 356 (1935); Codd in Bothalia
12: 10 (1976B); Codd in Fl. southern Afr. 30, 4: 201
(1985). – Type: Transvaal, Pretoria, Wilms 1151
(holotype, BM).
= O. natalensis Gürke, loc. cit. 82 (1898). – Type:
Natal, Glencoe, Medley Wood 4756 (lectotype, K!
chosen here), Syntypes: Kuntze s.n.; Coldstream,
Rehmann 6918 (K).
Syncolostemon punctatus (Codd) D.F. Otieno, comb.
nov. Hemizygia punctata Codd in Bothalia 12: 13
(1976b); Codd in Fl. southern Afr. 30, 4: 204 (1985).
– Type: Transvaal, 18 km SW of Lydenburg, Codd
8038 (holotype, PRE!; isotype NH!).
Syncolostemon ramosus (Codd) D.F. Otieno, comb.
nov. Hemizygia ramosa Codd in Bothalia 12: 14
(1976b); Codd in Fl. southern Afr. 30, 4: 204 (1985).
– Type: Natal, Lembobo Mts, near Mkuze, 7 Mar.
1966, Moll 3158 (holotype, PRE; isotypes, K!; NH!;
NU!).
Syncolostemon rehmannii (Gürke) D.F. Otieno, comb.
nov. Orthosiphon rehmannii Gürke in Bull. Herb.
Boiss. 6: 557 (1898); N.E. Br. in Fl. Cap. 5, 1: 251
(1910) Hemizygia rehmannii (Gürke) M. Ashby in
J. Bot., Lond. 73: 347 (1935); Codd in Bothalia 12:
8 (1976b); Codd in Fl. Southern Afr. 30, 4: 199
(1985). – Type: Transvaal, Houtbochsberg, Rehmann
6172 (holotype, Z; isotype, K).
Syncolostemon rugosifolius (M. Ashby) D.F. Otieno,
comb. nov. Hemizygia rugosifolia M. Ashby in J.
Bot. 73: 344 (1935); Codd in Bothalia 12: 4 (1976b);
Codd in Fl. southern Afr. 30, 4: 196 (1985). – Type:
Transvaal, The Downs, Junod 4342 (holotype,
PRE!).
Syncolostemon stalmansii (A.J. Paton & K. Balkwill)
D.F. Otieno, comb. nov. Hemizygia stalmansii A.J.
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
953
Paton & K. Balkwill in Kew Bull. 56: 492 (2001). –
Type: South Africa, Mpumalanga, Songimvelo
Game Reserve, Schoonoord Farm, 3.5 km from Ida
Doyer Picket, 2 Feb. 1999, K. Balkwill, M.-J.
Balkwill, M. Stalmans, D. Goyder & A. Paton 10891
(holotype J!; isotypes K, LISC).
Syncolostemon stenophyllus (Gürke) D.F. Otieno,
comb. nov. Orthosiphon stenophyllus Gürke in
Bot. Jahrb. 26: 84 (1898); N.E. Br. in Fl. Cap. 5, 1:
250 (1910) Hemizygia stenophylla (Gürke) M.
Ashby in J. Bot. 73: 347 (1935); Codd in Bothalia
12: 8 (1976b); Codd in Fl. southern Afr. 30, 4: 199
(1985). – Type: Natal, East Griqualand, near
Enyembe, Tyson 2137 (holotype, K!; isotype, PRE!),
Syntypes: Wood 966, 3197; Tyson s.n.; Mac Owen
1293.
Syncolostemon subvelutinus (Gürke) D.F. Otieno,
comb. nov. Orthosiphon subvelutinus Gürke in
Bot. Jahrb. 26: 80 (1898) Hemizygia subvelutina
(Gürke) M. Ashby in J. Bot. 73: 346 (1935); Codd in
Bothalia 12: 9 (1976b); Codd in Fl. southern Afr. 30,
4: 200 (1985). – Type: Transvaal, Lydenburg, near
Paarde Plaats, Wilms 1152 (holotype, BM!; isotype,
K).
=O. heterophyllus Gürke in Bot. Jahrb. 26: 82 (1898).
Syntypes: Transvaal, near Spitzkop, Wilms 1155 (K);
1148 (BM).
Syncolostemon teucriifolius (Hochst.) D.F. Otieno
comb. nov. Ocimum teucriifolium Hochst. in Flora
28: 66 (1844), Benth. in DC., Prodr. 12: 41 (1848)
Hemizygia teucriifolia (Hochst.) Briq. in Natürl.
Pflanzenfam. 4, 3a: 369 (1897); Codd in Bothalia
12: 9 (1976b); Codd in Fl. southern Afr. 30, 4: 200
(1985). – Type: Natal, Table Mt., Krauss 448 (holo-
type, BM; isotype, K!).
=Orthosiphon teucriifolius (Hochst.) N.E. Br. in Fl.
Cap. 5, 1: 254 (1910) Orthosiphon teucriifolius
var. galpinianus (Briq.) N.E. Br., loc. cit. 254 (1910).
Type: Transvaal, Barberton, Saddleback, Galpin
1217 (holotype, K!; isotypes, NH!, PRE!).
=Orthosiphon woodii Gürke in Bot. Jahrb. 26: 83
(1898). Type: Natal, Entumeni, Medley Wood sub
NH 783 (= Medley Wood 3964, holotype, K!; iso-
type, NH!).
=H. galpiniana Briq. in Bull. Herb. Boiss. ser. 2, 3:
993 (1903).
Syncolostemon transvaalensis (Schltr.) D.F. Otieno,
comb. nov. Orthosiphon transvaalensis Schltr. in
J. Bot., Lond. 35: 281 (1897) Hemizygia trans-
vaalensis (Schltr.) M.Ashby in J. Bot., Lond. 73: 349
(1935); Codd in Bothalia 12: 15 (1976b); Codd in Fl.
southern Afr. 30, 4: 205 (1985). – Type: Transvaal,
Barberton Galpin 468 (holotype, PRE; isotype,
SAM).
Syncolostemon thorncroftii (N.E. Br.) D.F. Otieno,
comb. nov. Orthosiphon thorncroftii N.E. Br. in Fl.
Cap. 5, 1: 246 (1910) Hemizygia thorncroftii (N.E.
Br.) M. Ashby in J. Bot. 73: 349 (1935); Codd in
Bothalia 12: 16 (1976b); Codd in Fl. southern Afr. 30,
4: 206 (1985). – Lectotype (Ashby 1936): Transvaal,
Barberton, Thorncroft 3123 (lectotype, K!).
ACKNOWLEDGEMENTS
The Herbarium Programme, School of Animal, Plant and
Environmental Sciences, University of the Witwatersrand, the
Lennox Boyd Memorial Trust and the Mellon Foundation sup-
ported this research.
LITERATURE CITED
Andreasen, K. & Bremer, B. 2000. Combined phylogenetic
analysis in the Rubiaceae – Ixoroideae: morphology,
nuclear and chloroplast DNA. Amer. J. Bot. 87: 1731–1748.
Ashby, M. 1935. The genus Hemizygia Briquet. J. Bot. 73:
312–318.
Baker, J. G. 1900. Labiatae. Pp. 332–526 in: Thiselton-Dyer,
W. T. (ed.), Flora of Tropical Africa. Lovell Reeve & Co,
London.
Bentham, G. 1837. Syncolostemon. Pp. 230 in: Meyer, E. (ed.),
Commentariorum de Plantis Africae Australis. Leopold
Voss, Leipzig.
Bentham, G. 1848. Ocimum. Pp. 31–44 in: Candolle, A. P, de
(ed.), Prodromus Systematis Naturalis. Treuttel et Würtz,
Paris.
Bremer, K. 1992. Ancestral Areas: a cladistic reinterpretation
of the center of origin concept. Syst. Biol. 41: 436–445.
Briggs, J. C. 1987. Biogeography and Plate Tectonics.
Elsevier, Amsterdam.
Briggs, J. C. 1989. The historic biogeography of India: isola-
tion or contact. Syst. Zool. 38: 322–332.
Briquet, J. 1897. Hemizygia. Pp. 368–369 in: Engler, A. &
Prantl, K. A. (eds.), Die Natürlichen Planzenfamilien, vol.
4, 3a. Wilhelm Engelmann, Leipzig.
Brown, N. E. 1910. Labiatae. Pp. 237 in: Thiselton-Dyer, W. T.
(ed.), Flora Capensis, vol. 5, 1. Lovell Reeve & Co,
London.
Buckland, A. & Bremer, K. 1998. To be or not to be—princi-
ples of classification and monotypic families. Taxon 47:
391–400.
Cantino, P. D., Harley, R. M. & Wagstaff, S. J. 1992. Genera
of Labiatae: status and classification. Pp. 511–513 in:
Harley, R. M. & Reynolds, T. (eds.), Advances in Labiatae
Science. Royal Botanic Gardens, Kew.
Carine, M. A. & Scotland, R. W. 2002. Classification of
Strobilanthinae (Acanthaceae): trying to classify the
unclassifiable? Taxon 51: 259–279.
Codd, L. E. 1976a. The genus Syncolostemon (Lamiaceae).
Bothalia 12: 21–28.
Codd, L. E. 1976b. The South African species of Hemizygia
(Lamiaceae). Bothalia 12: 1–20.
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
954
Codd, L. E. 1985. Syncolostemon and Hemizygia. Pp. 185–211
in: Leistner, O. A. (ed.), Flora of Southern Africa.
Botanical Research Institute, South Africa.
Codd, L. E. 1988. Resuscitation of Syncolostemon ramulosus
E. Mey. ex Benth. Bothalia 18: 92–93.
De Queiroz, A., Donoghue, M. J. & Junh Yong, K. 1995.
Separate versus combined analysis of phylogenetic evi-
dence. Ann. Rev. Ecol. Syst. 26: 657–681.
Donoghue, M. J. & Sanderson, M. J. 1992. The suitability of
molecular and morphological evidence in reconstructing
plant phylogeny. Pp. 340–368 in: Soltis, P. S., Soltis, D. E.
& Doyle J. J. (eds.), Molecular Systematics of Plants II.
Chapman & Hall, New York & London.
Doyle, J. J. & Doyle, J. L. 1987. A rapid DNA isolation pro-
cedure for small quantities of leaf tissue. Phytochem. Bull.
19: 11–15.
Eldenäs, P. K. & Linder, P. H. 2000. Congruence and com-
plementarity of morphological and trnL-trnF sequence
data and the phylogeny of the African Restionaceae. Syst.
Bot. 25: 692–707.
Farris, J. S. 1969. A successive approximations approach to
character weighting. Syst. Zool. 18: 274–385.
Farris, J. S. 1989. The Retention Index and the Rescaled
Consistency Index. Cladistics 5: 417–419.
Farris, J. S., Källersjö, M., Kluge, A. G. & Bult, C. 1995.
Constructing a significance test for incongruence. Syst.
Biol. 44: 570–572.
Felsenstein, J. 1985. Confidence limits on phylogenies: an
approach using bootstrap. Evolution 39: 783–791.
Fitch, W. M. 1971. Toward defining the course of evolution:
minimum change for a specific tree topology. Syst. Zool.
20: 406–416.
Goldblatt, P., Savolainen, V., Porteous, O., Sostaric, I.,
Powell, M., Reeves, G., Manning, J. C., Barraclough, T.
G. & Chase, M. W. 2002. Radiation in the Cape flora and
phylogeny of Peacock Irises Moraea (Iridaceae) based on
four plastid DNA regions. Molec. Phylog. Evol. 25:
341–361.
Holmgren, P. K., Holmgren, N. H. & Barnett, L. C. 1990.
Index Herbariorum, Part 1. The Herbaria of the World,
Ed. 8. New York Botanical Gardens, New York.
Huxley, C. R., Lock, J. M. & Cutler, D. F. (eds.) 1998.
Chorology, Taxonomy and Ecology of the Floras of Africa
and Madagascar. Royal Botanic Gardens, Kew.
Klompen, J. S. H., Black, W. C., Keirans, J. E. & Norris, D.
E. 2000. Systematics and biogeography of hard ticks, a
total evidence approach. Cladistics 16: 79–102.
Kluge, A. G. 1989. A concern for evidence and a phylogenetic
hypothesis of relationships among Epicrates (Boidae,
Serpentes). Syst. Zool. 38: 7–25.
Leroy, J.-F. 1978. Composition, origin and affinities of the
Madagascan vascular flora. Ann. Missouri Bot. Gard. 65:
535–589.
Moore, S. 1911. Labiatae. Bot. J. Linn. Soc. 40: 170–180.
Otieno, D. F., Balkwill, K. & Paton, A. J. 2006. A multivari-
ate analysis of morphological variation in the Hemizygia
bracteosa complex (Lamiaceae, Ocimeae). Pl. Syst. Evol.
261: 19–38.
Paton, A. 1997. Classification and species of Platostoma and its
relationship to Haumaniastrum. Kew Bull. 52: 257–292.
Paton, A. 1998. New records and new combinations in Hemi-
zygia and Syncolostemon. Kew Bull. 53: 483–485.
Paton, A. & Balkwill, K. 2001. Hemizygia stalmansii
(Labiatae), a new species from Mpumalanga, South Africa
and Swaziland. Kew Bull. 56: 491–496.
Paton, A., Harley, R. M. & Harley, M. M. 1999. Ocimum: an
overview of classification and relationships. Pp. 1–38 in:
Hiltunen, R. & Holm, Y. (eds.), The Genus Ocimum.
Medicinal and Aromatic Plants-Industrial Profiles.
Harwood Academic Publishers, Amsterdam.
Paton, A. & Hedge, I. C. 1999. Hemizygia. Pp. 108–111 in:
Hedge, I. C., Clement, R. A., Paton, A. J. & Phillipson, P.
B. (eds.), Flore de Madagascar et des Comores. Museum
National D’Histoire Naturelle, Paris.
Paton, A., Springate, D., Suddee, S., Otieno, D., Grayer,
R. J., Harley, M., Willis, F., Simmonds, M. S. J.,
Powell, M. & Savolainen, V. 2004. Phylogeny and evo-
lution of Basils and Allies (Ocimeae, Labiatae) based on
three Plastid DNA regions. Molec. Phylog. Evol. 31:
277–299.
Poe, S. & Wiens, J. J. 2000. Character selection and the
methodology of morphological phylogenetics. Pp. 20–36
in Wiens, J. J. (ed.), Phylogenetic Analysis of Morphol-
ogical data. Smithsonian Institution Press, Washington &
London.
Rabinowitz, P. D., Coffin, M. F. & Falvey, D. 1983. The sep-
aration of Madagascar and Africa. Science 220: 67–69.
Raven, P. H. & Axelrod, D. I. 1974. Angiosperm biogeogra-
phy and past continental movements. Ann. Missouri Bot.
Gard. 61: 539–661.
Reeves, G., Goldblatt, P., Chase, M. W., Rudall, P. J., Fay,
M. F., Cox, A. V., Lejeune, B. & Souza-Chies, T. 2001.
Molecular systematics of Iridaceae: evidence from four
plastid DNA regions. Amer. J. Bot. 88: 2074–2087.
Ryding, O. 1992. Pericarp structure and phylogeny within
Lamiaceae subfamily Nepetoideae tribe Ocimeae. Nord. J.
Bot. 12: 273–298.
Seelanan, T., Schnabel, A. & Wendel, J. F. 1997. Congruence
and consensus in the cotton tribe (Malvaceae). Syst. Bot.
22: 259–290.
Simmons, M. P. & Ochoterena, H. 2000. Gaps as characters
in sequence-based phylogenetic analyses. Syst. Biol. 49:
369–381.
Soltis, D. E., Soltis, P. S., Mort, M. E., Chase, M. W.,
Savolainen, V., Hoot, S. B. & Morton, C. M. 1998. In-
ferring complex phylogenies using parsimony: an empiri-
cal approach using three large DNA sets for angiosperms.
Syst. Biol. 47: 32–42.
Swofford, D. L. 2001. PAUP*. Phylogenetic Analysis Using
Parsimony (*and Other Methods), Version 4. Sinauer
Associates, Sunderland, Massachusetts.
Taberlet, P., Gielly, L., Paton, G. & Bonvet, J. 1991.
Universal primers for amplification of three non-coding
regions of chloroplast DNA. Pl. Molec. Biol. 17:
1105–1109.
Van der Bank, M., Fay, M. F. & Chase, M. W. 2002.
Molecular phylogenetics of Thymelaeaceae with particu-
lar reference to African and Australian genera. Taxon 51:
329–339.
Van Wyk, A. E. & Smith, G. F. 2001. Regions of Floristic
Endemism in Southern Africa. A Review with Emphasis on
Succulents. Umdaus Press, Pretoria.
Wagstaff, S. J., Olmstead, R. G. & Cantino, P. D. 1995.
Parsimony analysis of cpDNA restriction site variation in
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
955
subfamily Nepetoideae (Labiatae). Amer. J. Bot. 82:
886–892.
Weijermars, R. 1989. Global tectonics since the break up of
Pangea 180 million years ago: evolution maps and lithos-
pheric budget. Earth-Science Rev. 26: 113–162.
White, F. 1983. The Vegetation of Africa: A Descriptive
Memoir to accompany the UNESCO/AETFAT/UNSO
Vegetation Map of Africa. Natural Resources Research
XX. Unesco, Paris.
Wiens, J. J. 1998. Combining data sets with different phylo-
genetic histories. Syst. Biol. 47: 568–581.
Williams, C. A. 1986. An oceanwide view of paleogene plate
tectonic events. Palaeogeog. Palaeoclimatol. Paleoecol.
57: 3–25.
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
956
Appendix 2. Characters used in morphological and combined data analysis of Hemizygia and Syncolostemon.
1. Leaf fascicles arising from axils of leaves on branches: absent (0); present (1). 2. Leaf attachment: petiolate (0); sessile (1). 3. Leaf mar-
gin: flat (0); revolute (1). 4. Sessile glands: present on both abaxial and adaxial surfaces (0); present only on adaxial surface (1); absent
(2). 5. Sessile glands: raised (0); sunken (1). 6. Simple hairs: present (0); absent (1). 7. Dendroid hairs: absent (0); present (1). 8. Biramous
hairs: absent (0); present (1). 9. Glandular hairs: present (0); absent (1). 10. Orientation of leaf trichomes: upright (0); appressed (1). 11.
Inflorescence: terminal on short or long lateral shoots (0); simple but sometimes with one or two pairs of branches at the base (1); com-
pound with branches lax or crowded (2). 12. Coma formed by coloured terminal bracts: absent (0); present (1). 13. Flowers per cyme:
variable 1–8 (0); always 1 (1). 14. Bracts subtending cymes: persistent (0); caducous (1). 15. Shape of fruiting calyx: tubular (0); cam-
panulate (1); urceolate (2). 16. Calyx tube throat glandular: pubescent (0); covered with sparse to dense eglandular hairs interspersed with
short or long glandular hairs (1); setose (2); villous (3); glabrous (4). 17. Constriction at the distal end of fruiting calyx: absent (0); pres-
ent (1). 18. Anterior lobes of calyx: convergent (0); divergent (1). 19. Depth of sinus between lower lobes of anterior calyx lip: < that
between anterior and lateral lobes (0); > that between anterior and lateral lobes (1); equal to that between anterior and lateral lobes (2).
20. Posterior calyx lobe: decurrent on the tube (0); non-decurrent (1). 21. Shape of posterior calyx lobe: ovate to obovate (0); elliptic to
broadly elliptic (1); deltoid to deltoid-subulate (2). 22. Posterior calyx lobe: strongly raised (0); not raised at all or very weakly raised (1).
23. Corolla tube: straight (0); geniculate (1); slightly curved upwards (2); curved downwards (3). 24. Corolla: dilating in varying degrees
from the base towards a wide mouth (0); dilating from the base but with a distinct constriction at the mouth (1); parallel-sided for a short
distance from the base or up to the middle, then dilating towards a slightly constricted mouth (2); parallel-sided for a short distance from
the base or up to the middle then dilating towards a gaping mouth (3); parallel-sided from the base, through the middle and only dilating
near the mouth (4); parallel-sided from the base, through the middle up to the mouth (5). 25. Corolla tube interior: sparsely to densely
pubescent (0); setose (1); with a ring of hairs (2); glabrous (3). 26. Anterior corolla lip: bilobed with a deep sinus (0); one-lobed (1). 27.
Anterior stamen: free (0); fused (1). 28. Posterior stamens inserted: at or above the middle of corolla tube but below the mouth (0); below
the middle of corolla tube (1). 29. Posterior stamens: included (0); not exserted beyond the lower corolla lip but occasionally exceeding
the latter by a distance less than or equal to half its length (1); always exserted beyond the lower corolla lip (2). 30. Hairy appendage near
the base of posterior stamens: absent (0); present (1). 31. Anterior stamens: included (0); not exserted beyond the lower corolla lip but
occasionally exceeding the latter by a distance less than or equal to half its length (1); always exserted beyond the lower corolla lip (2).
32. Fusion of anterior stamens: always extending above the middle of the filaments from the base (0); never extending above the middle
of the filaments from the base (1). 33. Indumentum on anterior stamens: covering whole stamen (0); only at the base (1); lacking (2). 34.
Stigma: bifid (0); clavate (1). 35. Style shield: present (0); absent (1). 36. Length of nutlet: > 1 mm (0); > 0.5 mm but to 1 mm (1). 37.
Color of nutlet: brown (0); brownish-black to black (1). 38. Apex of nutlet: truncate (0); obtuse to rounded (1). 39. Hairs on apex of nut-
let: present (0); absent (1). 40. Nutlet cross-section: obtriangular (0); elliptic to obovate (1). 41. Frill at the base of nutlet: absent (0); pres-
ent (1). 42. Vein on nutlet: none (0); unbranched (1); branched (2). 43. Lobes of disk: unequal (0); equal (1). 44. Apex of stylopodium:
not lobed (0); lobed (1). 45. Persistent remains of the style at the apex of the stylopodium: raised (0); sunken (1). 46. Buttresses on sty-
lopodium: absent (0); present (1). 47. Membranous structures hanging outwards from placental scar on stylopodium: present (0); absent
(1). 48. Length of stylopodium from the position where the membranous structures are attached up to the apex: short (0); long (1).
Appendix 1. List of plant samples with GenBank/EMBL accession numbers and voucher information.
Species, Collection number & voucher location, Genbank/EBI accessions (rps16, trnL-F).
Endostemon obtusifolius (E. Mey. ex Benth) N.E. Br., Goyder & Paton 4105 (K), No sequence, AJ505442; Hemizygia albiflora (N.E. Br.)
M. Ashby, Otieno 66 (J), AJ563764, AJ563783; H. foliosa S. Moore, Otieno 45 (J), AJ563765, AJ563784; H. incana Codd, Otieno 75 (J),
AJ563766, AJ563785; H. modesta Codd, Otieno 120 (J), AJ563767, AJ563786; H. obermeyerae M. Ashby, Otieno 13 (J), AJ563768,
AJ563787; H. parvifolia Codd, Otieno 71 (J), AJ563769, AJ563788; H. persimilis (N.E. Br.) M. Ashby, Otieno 76 (J), AJ563770,
AJ563789; H. pretoriae (Gürke) M. Ashby, Otieno 50 (J), AJ563771, AJ563790; H. punctata Codd, Otieno 72 (J), AJ563772, AJ563791;
H. stalmansii A.J. Paton & K. Balkwill, Otieno 111 (J), AJ563773, AJ563792; H. subvelutina (Gürke) M. Ashby, Otieno 69 (J), AJ563774,
AJ563793; H. transvaalensis (Schltr.) M. Ashby, Otieno 56 (J), AJ563775, AJ563794; H. teucriifolia (Hochst.) Briq., Otieno 64 (J),
AJ563776, AJ563795; Ocimum labiatum (N.E. Br.) A.J. Paton, Balkwill et al. 10848 (K), AJ505366, AJ505471; O. serratum (Schltr.) A.J.
Paton, Balkwill et al. 10861 (J, K), AJ505357, AJ505472; O. tenuiflorum L., Suddee et al. 893 (BKF, K, TCD), AJ505358, AJ505473;
Syncolostemon argenteus N.E. Br., Otieno 93 (J), AJ563777, AJ563796; S. comptonii Codd, McCallum 916 (J), AJ563778, AJ563797; S.
flabellifolius (S. Moore) A.J. Paton, Linder 3981 (J), AJ563779, AJ563798; S. macranthus (Gürke) M. Ashby, Otieno 109 (J), AJ563780,
AJ563801; S. parviflorus E. Mey. ex Benth., Otieno 84 (J), AJ563781, AJ563800; S. rotundifolius E. Mey. ex Benth., Otieno 82 (J),
AJ563782, AJ563799.
Otieno & al. • Hemizygia and Syncolostemon55 (4) • November 2006: 941958
957
Appendix 3. Data matrix for cladistic analysis of Hemizygia and Syncolostemon. Outgroup species is denoted with *.
Character states are scored 0 to 5, “?” = missing data, “-” = inapplicable data, a = 0 & 1, b = 0 & 2, c = 0 & 3, d = 1 & 2.
Characters
0000000001 1111111112 2222222223 3333333334 44444444
Taxon
1234567890 1234567890 1234567890 1234567890 12345678
Endostemon obtusifolius*
0000000000 0000000000 0000000000 0-00000000 00000000
Hemizygia albiflora
1012-00011 1001010020 0032011120 2020100111 01000101
H. bolusii
0000100000 10010????0 003001a120 2010100111 010?????
H. bractoesa 1
0100000000 1101010020 0033011120 2020100111 02011101
H. bracteosa 2
0100000000 1101010020 0033011120 2020100111 01011101
H. bracteosa 3
0111000000 1101010020 0034011120 201?100111 01011101
H. canescens
0002-00000 1001010020 0024011120 2020101111 01011101
H. cinerea
0011011000 1001010000 0033011120 2010100111 01000101
H. comosa
0110000000 1121010010 0035311010 1120100111 01011001
H. elliottii
0000011000 1101010020 0030011110 1121100111 01011101
H. floccosa
0002-11a00 1011010000 0030211111 1121100111 01011101
H. foliosa
0a00000000 1111010010 003c011120 2010100111 01000101
H. gerrardii
0000011000 1101010020 00c4011020 2020100111 01000101
H. incana
0111011000 1001110000 0033011120 20d0100111 01000101
H. linearis
0102-00000 1001010020 0024011120 2020101111 01011101
H. macrophylla
0000111000 2001221111 0133211120 2020100111 0100011-
H. madagascariensis
00000000a0 1001010010 0030011110 1020100111 01000001
H. modesta
0000100000 10110100a0 0030011120 20d0100111 01000101
H. obermeyerae
0010011000 1101010000 0033011120 20d0100111 01000101
H. oritrephes
0112-01a00 1001010000 0033211110 1020100111 01000101
H. ornata
0100000000 1101010000 0013011120 2020100111 01000101
H. parvifolia
0011011000 1011010000 0033011120 2010100111 01000101
H. persimilis
0100100000 1100010010 00c0011100 1111100111 01011101
H. petiolata
0002-00000 1101010000 0024011120 2020101111 01011101
H. petrensis
0a02-00000 10010100b0 0024011120 2020101111 01011101
H. pretoriae 1
0000100000 1000010010 0003011000 1020100111 01000101
H. pretoriae 2
0000101100 10000100a0 0003011000 1020100111 01000101
H. punctata
0000100000 1011010010 0030011120 2020100111 01000101
H. ramosa
0000100000 1101010010 0033011020 2a201?0111 01000101
H. rehmannii
0110011a10 1001010000 0033011120 20d0100111 01000101
H. rugosifolia
0010011100 1101011010 00c3011120 20201?0111 010?????
H. sp. nov.
0100100000 1101010010 0020011110 1020100111 01000101
H. stalmansii
0010000000 21010200a0 0033011120 2020100111 01000101
H. stenophylla
0112-11000 1101010020 0033011120 20d0100111 01000101
H. subvelutina
1011011010 1000010020 0031211110 1a20100111 01000101
H. teucriifolia
11100a1a00 1000010020 0031211110 1020100111 01000101
H. thorncroftii
0100a00000 1101010010 0033011120 2010100111 01000101
H. transvaalensis
0aa0000000 1101010010 00330111d0 d020100111 01000101
H. welwitschii
0010000000 1101010000 0033011120 2a10100111 01011101
Ocimum labiatum
0000000010 1001040000 0030310110 1-20100111 00000000
O. serratum
0011000010 1001040000 0030010110 1-20100111 00000000
O. tenuiflorum
0000100010 2001140000 0000310121 2-20110111 00111101
Syncolostemon argenteus
1100100011 2011231121 2130211120 2020100111 1100011-
S. comptonii
1100100010 2011221?11 2130211120 2020100111 1100011-
S. concinnus
1100111010 2011221111 2130211120 2010100111 1100011-
S. densiflorus
1000100010 2001021111 2130011120 2020100111 0100011-
S. eriocephalus
1110100011 0021130111 2130211120 2010100111 0100011-
S. flabellifolius
0000000011 100112011a 0030211110 1001100111 0100011-
S. latidens
0000000000 2001020011 1133011120 2020100111 0100011-
S. macranthus
0000100000 2001020011 1133011020 20d0100111 0100011-
S. parviflorus
1a00100010 2011221111 2130b11120 2020100111 1100011-
S. ramulosus
1000100010 2001021111 2133011120 2020100111 0100011-
S. rotundifolius
0000100010 2011021111 2133011a20 20d0100111 0100011-
S. sp. nov.
1100100010 2011221111 2134111110 1020100111 1100011-
Otieno & al. • Hemizygia and Syncolostemon 55 (4) • November 2006: 941958
958
Appendix 4. Statistics of morphological characters mapped to the phylogeny in MP analysis.
Character CI RI
1. Leaf fascicles arising from axils of leaves on branches 0.33 0.80
2. Leaf attachment 0.10 0.52
3. Leaf margin 0.16 0.68
4. Sessile glands 0.22 0.41
5. Disposition of sessile glands 0.16 0.73
6. Simple hairs 0.16 0.58
7. Dendroid hairs 0.16 0.66
8. Biramous hairs 0.50 0.00
9. Glandular hairs 0.16 0.68
10. Orientation of leaf trichomes 0.33 0.33
11. Inflorescence branching 0.50 0.84
12. Coma formed by coloured terminal bracts 0.12 0.61
13. Flowers per cyme 0.25 0.45
14. Bracts subtending cymes 0.25 0.40
15. Shape of fruiting calyx 0.33 0.50
16. Calyx tube throat 0.80 0.92
17. Constriction at distal end of fruiting calyx 0.33 0.77
18. Anterior lobe of calyx 0.50 0.88
19. Depth of sinus between lower lobes of anterior calyx lip 0.16 0.60
20. Attachment of posterior calyx lobe 1.00 1.00
21. Shape of posterior calyx lobe 0.66 0.89
22. Orientation of posterior calyx lobe 1.00 1.00
23. Corolla tube curvature 0.50 0.57
24. Corolla expansion 0.62 0.87
25. Corolla tube interior 0.33 0.45
26. Anterior corolla lip 1.00 0.00
27. Fusion of stamens 1.00 1.00
28. Insertion position of posterior stamens 0.20 0.33
29. Exsertion of posterior stamens 0.28 0.64
30. Hairy appendage near the base of posterior stamens 0.50 0.00
31. Exsertion of anterior stamens 0.33 0.71
32. Position at which filaments of anterior stamens fused 1.00 1.00
33. Indumentum on anterior stamens 0.22 0.30
34. Stigma 0.50 0.66
35. Style shield 1.00 0.00
36. Length of nutlet 1.00 0.00
37. Colour of nutlet 1.00 1.00
38. Apex of nutlet 1.00 0.00
39. Hairs on nutlet apex 1.00 00.0
40. Nutlet cross-section 1.00 0.00
41. Frill at the base of the nutlet 0.50 0.75
42. Nutlet vein 1.00 1.00
43. Disk lobing 1.00 1.00
44. Apex of stylopodium 0.33 0.83
45. Persistent remains of the style at the apex of the stylopodium 0.33 0.83
46. Buttresses on stylopodium 0.33 0.50
47. Membranous structures hanging outwards from placental scar on stylopodium 0.50 0.91
48. Length of stylopodium from the position where the membranous structures are attached up to the apex 1.00 1.00
... Codd (1985) recognised nine species of Syncolostemon in his treatment of the family for the Flora of southern Africa, but this number has increased with the inclusion of the 33 species of Hemizygia (Benth.) Briq. in Syncolostemon (Otieno et al. 2006). Syncolostemon is characterised by the presence of a bilabiate or subequally 5-toothed calyx, the uppermost tooth larger than the remaining 4, broadly ovate to subrotund, tube campanulate, cylindrical, 2 lower teeth subulate to spinescent. ...
... The genus currently includes 44 species from Africa, with the majority occurring in southern Africa, and one species from outside Africa (Syncolostemon comosus (Wight ex Benth.) D.F.Otieno from India) and one species from Madagascar (Syncolostemon madagascariensis (A.J.Paton & Hedge) D.F.Otieno) (Otieno et al. 2006). Syncolostemon was described to accommodate material collected by J.F. Drège, and the species were separated into two groups. ...
... and S. parviflorus), and Group 2 included just the two original species (S. densiflorus and S. rotundifolius). Otieno et al. (2006) concluded that Syncolostemon s.l. is not found throughout the Afroalpine and Afromontane Regional Centre of Endemism. This is an archipelago-like area that ranges from the southern Drakensberg of South Africa through the high mountains of East Africa and north to the Simiana Mountains of Ethiopia. ...
Article
Full-text available
A new species of Syncolostemon E.Mey. ex Benth. (Lamiaceae) from north-western KwaZulu-Natal (Utrecht), South Africa, is described. Syncolostemon aurulentus is notable in having conspicuous glittering golden glands on the leaves, bracts and calyces, which are absent in other species of the genus.
... Study taxa were selected from a small group of ten southern African species that, until recently, were considered to make up the entire genus Syncolostemon (Codd). A recent analysis has shown that this group forms a clade that is embedded within the more wide-ranging genus Hemizygia ( Otieno et al. 2006) As a result, Syncolostemon (the name that has priority) has been broadened to include all the former species of Hemizygia and now consists of c. 44 species ( Otieno et al. 2006). Syncolostemon s.l. is characterized by partially fused filaments of the anterior stamens. ...
... Study taxa were selected from a small group of ten southern African species that, until recently, were considered to make up the entire genus Syncolostemon (Codd). A recent analysis has shown that this group forms a clade that is embedded within the more wide-ranging genus Hemizygia ( Otieno et al. 2006) As a result, Syncolostemon (the name that has priority) has been broadened to include all the former species of Hemizygia and now consists of c. 44 species ( Otieno et al. 2006). Syncolostemon s.l. is characterized by partially fused filaments of the anterior stamens. ...
... Long-proboscid flies and birds appeared to be the most effective pollinators of the long-tubed species from assessment of the amount and purity (in terms of proportion The three treatments applied to bagged flowers were cross-pollination (Cross), self-pollination (Self) and no manipulation (Cont) of Syncolostemon pollen) of their pollen loads and visitation frequency ( Table 2). The phylogenetic analysis of Otieno et al. (2006) suggests that the direction of corolla tube evolution in Syncolostemon was from short to long corollas in the study group (probably representing a shift from bee to long-proboscid fly and bird-pollination), while the shorttubed species S. parviflorus and S. ramulosus which are nested within the long-tubed clade appear to represent reversals to bee pollination. The alternative possibility, that long-tubed flowers evolved several times in parallel from short-tubed flowers, is less parsimonious on account of the greater number of steps required. ...
Article
Differentiation in floral traits among clusters of related species may reflect a process of pollinator-driven evolution. Pollination systems in the morphologically diverse southern African genus Syncolostemon (Lamiaceae) were investigated by means of field observations of floral visitors and analysis of their pollen loads. Among the five study species, those with short corolla tubes (S. parviflorus, S. ramulosus) were pollinated solely by bees, while those with long corolla tubes were pollinated by a broader array of visitors, primarily long-proboscid flies in S. rotundifolius and S. macranthus and sunbirds in S. densiflorus. The predominately insect-pollinated taxa have lax inflorescences, but S. densiflorus has a compact terminal inflorescence, which facilitates feeding by sunbirds from a single perching position. Experimental hand-pollinations involving three taxa (S. macranthus, S. densiflorus, and S. rotundifolius) showed that these possess a genetic self-incompatibility system. Production of fruits and seeds per fruit was pollen-limited in S. densiflorus and S. rotundifolius, but not in S. macranthus.
... 2.3 µL; Ford & Johnson, 2008) and pink flower colour are more consistent with the floral features of long-proboscid fly-pollinated plants (Baumberger, 1987;Goldblatt & Manning, 2000;Rodríguez-Gironés & Santamaría, 2004;Johnson & Nicolson, 2008;but see Azpeitia & Lara, 2006;Wester & Claßen-Bockhoff, 2011;Vogel, 2012). It is possible that these characters are a phylogenetic 'carry-over' as the ancestors of S. densiflorus were probably pollinated by insects (Ford & Johnson, 2008; see also Otieno et al., 2006). It is also possible that as secondary pollinators, long-proboscid flies and other insects have contributed to the selective maintenance of these traits in S. densiflorus. ...
Article
Flowers visited by a broad assemblage of animals may nevertheless be considered to have a specialized pollination system if pollen deposition is effected mainly by a subset of these visitors. The importance of a particular animal group for pollination derives mainly from the relative rate of floral visitation and per-visit effectiveness for pollen deposition. We assessed the relative importance of birds and insects for pollination of the African shrub Syncolostemon densiflorus (Lamiaceae). Flower visitors were observed to evaluate their visitation rates. Virgin flowers were exposed to single visits by nine animal visitor groups to evaluate their effectiveness at depositing pollen on stigmas. We selectively excluded birds and larger insects from flowers to assess the contribution of smaller visitors to pollen deposition and seed set. Sunbirds and insects visited flowers at a similar frequency, but on a per-visit basis, sunbirds deposited more pollen on stigmas than did any of the insect groups, including honeybees, long-proboscid flies, butterflies and day-flying hawkmoths. Selective exclusion of birds and larger insects resulted in a large decline in pollen deposition on stigmas and seed production. These results show that S. densiflorus is pollinated mainly by sunbirds, despite its attraction of a broad assemblage of insect flower visitors. These findings highlight the importance of per-visit pollination effectiveness and selective exclusion for understanding the ecological importance of different flower visitors.
... A re-assessment of the relationship between Syncolostemon and Hemizygia based on molecular and morphological data (Otieno et al. 2006b), showed that neither genus as previously circumscribed, was monophyletic, and thus led to the merging of Hemizygia into Syncolostemon, the earliest name. Members of the genus, as newly circumscribed, are characterized by fused anterior stamens, a feature that sets them apart from other genera in the tribe Ocimeae (Paton 1998). ...
... A re-assessment of the relationship between Syncolostemon and Hemizygia based on molecular and mor­ phological data (Otieno et al. 2006b), showed that neither genus as previously circumscribed, was monophyletic , and thus led to the merging of Hemizygia into Syncolostemon, the earliest name. Members of the genus, as newly circumscribed, are characterized by fused ante­ rior stamens, a feature that sets them apart from other genera in the tribe Ocimeae (Paton 1998). ...
Article
Full-text available
DROSERA ERICGREENII. A NEW SPECIES FROM THE FYNBOS OF SOUTH AFRICA
... Briq. (Otieno et al. 2006)) (Paton et al. 2004). Orthosiphon differs from both these genera by having short rounded or pointed lobes of the style apex, rather than diverging subulate lobes. ...
Article
Full-text available
A new species, Orthosiphon americanus Harley & A. J. Paton (Lamiaceae), is described. This represents the first known occurrence of the genus in the New World. The species’ relationships and phytogeography are discussed.
Article
In this worldwide revision of Syncolostemon sensu lato (including Hemizygia), 51 species are recognised, with three divided into varieties. Forty-nine species occur on mainland Africa, one on Madagascar and one in India. Descriptions, synonymy, habitat and conservation information are provided for each species and an identification key presented. Three new species, S. rectiflorus K.Balkwill, S. qudeniensis D.F.Otieno and S. ngwenyi K.Balkwill, and one new variety, S. pretoriae var. longitubus K.Balkwill are described and two new combinations, S. heterophyllus (Gürke) K.Balkwill and S. muddii (N.E.Br.) K.Balkwill are made. Bouetia ocimoides A.Chev. that had previously been synonymised under H. bracteosa (Benth.) Briq. [= S. bracteosus (Benth.) D.F.Otieno] is now recognised as a variety of that species.
Article
Full-text available
The Lamiaceae is undoubtedly an important plant family, having a rich history of use that spans the globe with many species being used in folk medicine and modern industries alike. Their ability to produce aromatic volatile oils has made them valuable sources of materials in the cosmetic, culinary, and pharmaceutical industries. A thorough account of the taxonomic diversity, chemistry and ethnobotany is lacking for southern African Lamiaceae, which feature some of the region’s most notable medicinal and edible plant species. We provide a comprehensive insight into the Lamiaceae flora of southern Africa, comprising 297 species in 42 genera, 105 of which are endemic to the subcontinent. We further explore the medicinal and traditional uses, where all genera with documented uses are covered for the region. A broad review of the chemistry of southern African Lamiaceae is presented, noting that only 101 species (34%) have been investigated chemically (either their volatile oils or phytochemical characterization of secondary metabolites), thus presenting many and varied opportunities for further studies. The main aim of our study was therefore to present an up-to-date account of the botany, chemistry and traditional uses of the family in southern Africa, and to identify obvious knowledge gaps.
Article
Full-text available
Parsimony analysis of cpDNA restriction site variation supports monophyly of subfamily Nepetoideae. However, a close relationship among Nepetoideae and other gynobasic-styled Labiatae is not supported, indicating that a gynobasic style has evolved independently in at least two clades of Labiatae. The inferred relationships are congruent with the classification of Cantino, Harley, and Wagstaff (1992, Advances in labiate sciences, 27–37, Royal Botanic Gardens, Kew) but conflict to varying degrees with traditional classifications. Monophyly of four tribes of Nepetoideae also is supported.
Article
Full-text available
The South African species of Hemizygia are reviewed and 28 species are recognized, including the following new names: H. macrophylla (Guerke) Codd (=Syncolostemon macrophyllus Guerke), H. pretoriae Guerke var. heterotricha Codd, H. cinerea Codd, H. incana Codd, H. modesta Codd, H. parvifolia Codd, H. punctata Codd and H. ramosa Codd.
Article
Full-text available
The genus Syncolostemon E. Mey. ex Benth. is revised and its relationship with Hemizygia (Benth.) Briq.f is discussed. A new species, S. comptonii Codd is described and the following new combinations are made: S. parviflorus var. lanceolatus (Guerke) Codd (= S . lanceolatus Guerke) and S. latidens (N.E. Br.) Codd ( = Orthosiphon latidens N.E. Br.).
Article
The recently-developed statistical method known as the "bootstrap" can be used to place confidence intervals on phylogenies. It involves resampling points from one's own data, with replacement, to create a series of bootstrap samples of the same size as the original data. Each of these is analyzed, and the variation among the resulting estimates taken to indicate the size of the error involved in making estimates from the original data. In the case of phylogenies, it is argued that the proper method of resampling is to keep all of the original species while sampling characters with replacement, under the assumption that the characters have been independently drawn by the systematist and have evolved independently. Majority-rule consensus trees can be used to construct a phylogeny showing all of the inferred monophyletic groups that occurred in a majority of the bootstrap samples. If a group shows up 95% of the time or more, the evidence for it is taken to be statistically significant. Existing computer programs can be used to analyze different bootstrap samples by using weights on the characters, the weight of a character being how many times it was drawn in bootstrap sampling. When all characters are perfectly compatible, as envisioned by Hennig, bootstrap sampling becomes unnecessary; the bootstrap method would show significant evidence for a group if it is defined by three or more characters.
Chapter
Renewed interest in phylogenies over the last few decades coincides with a growing sense that it will actually be possible to obtain an accurate picture of evolutionary history. Indeed, the prospects of retrieving phylogeny now seem better than ever, owing to basic theoretical advance (due mainly to Hennig, 1966), the availability of computer programs that can handle large data sets, and the accessibility of new sources of evidence, especially molecular characters.
Article
The generic delimitation of Platostoma P. Beauv., Haumaniastrum P.A. Duvign. and Plancke and previously recognized allied genera is examined on a global basis for the first time this century. A parsimony analysis was carried out and the generic concept applied is discussed. Platostoma and Haumaniastrum are recognized here with Acrocephalus, Geniosporum, Ceratanthus, Octomeron, Mesona, Nosema and Limniboza being placed in the synonomy of Platostoma. The previously recognized Acrocephalus group of genera is renamed the Platostoma group. Benguellia is considered to be more closely related to Orthosiphon than to the Platostoma group. Platostoma is a paraphyletic genus due to the exclusion of Haumaniastrum. Platostoma is divided into three subgenera: subgen. Acrocephalus is broadly equivalent to Asiatic species of Geniosporum and Acrocephalus, but with the inclusion of some African and Madagascan species: subgen. Octomeron is monotypic: subgen. Platostoma is divided into four sections, sect. Platostoma, sect. Ceratanthus, sect. Mesona and sect. Limniboza. Platostoma is found in tropical Africa, Madagascar, tropical Asia and N Queensland. The genus has not been fully revised, but 45 species are recognized with 4 described for the first time. A key to the infrageneric taxa is provided and the recognized species are listed with their types, synonomy and distribution. The mainly African genus Haumaniastrum is found to be more closely related to the African Platostoma sect. Limniboza rather than the Asiatic sect. Acrocephalus as had been previously proposed.