05Mackinder
11/2/10
3:54 pm
Page 43
Mackinder, B. A., Wieringa, J. J., Lunenburg, I. & Banks, H. (2010). Clarifying the generic limits
of Talbotiella and Hymenostegia (Detarieae, Caesalpinioideae, Leguminoseae). In: X. van der Burgt,
J. van der Maesen & J.-M. Onana (eds), Systematics and Conservation of African Plants, pp. 43–56.
Royal Botanic Gardens, Kew.
CLARIFYING THE GENERIC LIMITS
OF TALBOTIELLA AND HYMENOSTEGIA
(DETARIEAE, CAESALPINIOIDEAE, LEGUMINOSAE)
BARBARA A. MACKINDER1, JAN J. WIERINGA2, ILONA LUNENBURG2
& HANNAH BANKS1
1Royal
Botanic Gardens, Kew, Richmond, Surrey, TW9 3AB, UK;
b.mackinder@kew.org
2 Wageningen University, Biosystematics Group, National Herbarium of the
Netherlands, Wageningen University Branch, Generaal Foulkesweg 37,
6703 BL Wageningen, the Netherlands
Abstract
The formal description of four species of Cameroonian forest legume trees new to
science has been hampered by uncertainty as to whether their correct generic
placement is within Hymenostegia Harms or Talbotiella Baker f. As there has been doubt
as to whether these two genera differ from one another, an investigation was
undertaken so that the new species could be correctly assigned to genus. Using
morphological, molecular and pollen data, our study supports the recognition of
Hymenostegia and Talbotiella as distinct genera, consequently the new species are
correctly placed in Talbotiella. In addition, our data reveal the extensive heterogeneous
nature of Hymenostegia as currently circumscribed and the need to transfer H. breteleri
to Talbotiella.
Résumé
Clarification des limites génériques de Talbotiella et Hymenostegia (Detarieae,
Caesalpinioideae, Leguminosae). La description formelle de quatre espèces nouvelles
à la science des arbres camerounais des Légumineuses de forêt a été entravée par
l’incertitude de savoir si leur placement générique correct est chez Hymenostegia
Harms ou Talbotiella Baker f. Car ayant douté de savoir si ces deux genres sont
différents l’un de l’autre, une recherche a été entreprise de sorte que les nouvelles
espèces aient pu être correctement assignées au genre correspondant. Utilisant les
données morphologiques, moléculaires et de pollen, notre étude confirme que
Hymenostegia et Talbotiella sont distincts et que les nouvelles espèces sont correctement
placées dans Talbotiella. En outre, nos données indiquent la nature hétérogène
étendue de Hymenostegia comme actuellement circonscrit et la nécessité de transférer
H. breteleri à Talbotiella.
Key words: Africa, Fabaceae, Leguminosae, morphology, palynology, phylogeny, pollen.
43
05Mackinder
11/2/10
3:54 pm
Page 44
Systematics and Conservation of African Plants
1 Introduction
This study was prompted by the need to describe four Cameroonian endemic legume
tree species as new to science. Prior to this study, a major impediment to their
description was uncertainty as to their correct generic placement. All the undescribed
species have paripinnate, multijugate leaves, apetalous flowers and belong to tribe
Detarieae. Their correct generic assignment could be determined either to
Hymenostegia Harms or to Talbotiella Baker f. but the boundary between the two genera
was unclear. Morphological, nucleotide sequence and pollen surface characters were
appraised in an attempt to resolve one fundamental question: Is Talbotiella distinct from
Hymenostegia or are they congeneric?
In 1897, Harms established the genus Hymenostegia to accommodate three species
which he transferred from Cynometra. During the next 70 or so years a further 24 species
of Hymenostegia were described, although many were subsequently transferred to other
genera. Currently, Hymenostegia is a poorly understood genus of 16 forest tree species
from West and West-Central Africa (Mackinder, 2005).
In 1914 E. G. Baker described the monotypic genus Talbotiella to accommodate a
species of caesalpinioid legume new to science, T. eketensis, from SE Nigeria (Baker,
1914). In 1928 Hutchinson and Greenway added a second species, T. gentii from Ghana
(Hutchinson & Dalziel, 1928) and in the following year, Baker made a third addition to
the genus with the publication of T. batesii from Cameroon (Baker, 1929). No further
species have been described since then but there have been persistent doubts about the
generic status of Talbotiella with respect to Hymenostegia (Léonard, 1951, 1957; Cowan &
Polhill, 1981).
2 Léonard’s postulate of a single evolutionary series
Talbotiella is characterised by the combination of persistent imbricate bud scales,
persistent narrow petaloid bracteoles and the absence of petals, but many Hymenostegia
species possess one or both of the first two characters (Table 1).
TABLE 1. Character combination on which the genus Talbotiella was based and the
comparative character states found in Hymenostegia.
Character
Persistent imbricate bud scales
Persistent narrow petaloid bracteoles
Petals
Talbotiella
Hymenostegia
always present
always present
0
often present
sometimes present
(1) 2 – 5
In 1951 on consecutive pages, Léonard published accounts of Hymenostegia (Léonard,
1951a) and Talbotiella (Léonard, 1951b). He noted that the species of Hymenostegia
presented a progressive reduction of petals from 5 to 2(1) and postulated that the
apetalous Talbotiella may merely be the end of an evolutionary series that began in
Hymenostegia. Léonard suggested that Talbotiella might be better treated as a section of
Hymenostegia, but he stopped short of formalising the transfer until more material of
Talbotiella became available for study.
44
05Mackinder
11/2/10
3:54 pm
Page 45
Clarifying the generic limits of Talbotiella and Hymenostegia
In this paper we considered Léonard’s postulate in the light of new data, specifically
vegetative characters, nucleotide sequence data and pollen surface characters, and we
re-examined floral characters. If new data confirm Léonard’s postulate, then Talbotiella
must be considered congeneric with Hymenostegia. In order not to render Hymenostegia
paraphyletic the three described species of Talbotiella should be transferred there and
the four species new to science must be described in Hymenostegia, resulting in its
expansion to 23 species.
If, however, new data refute Léonard’s postulate, then Talbotiella should be upheld as
a distinct genus within which the four species new to science will be described bringing
the total to seven species. Hymenostegia would thus remain a genus of 16 species.
3 Evidence from vegetative morphology
Vegetative characters examined in the study were those which had previously shown
utility, either singly, or more frequently, in combination, in the recognition of either
generic or species limits in other members of Detarieae. The 23 vegetative
morphological characters surveyed are listed in Table 2. Characters were scored from
selected herbarium specimens on loan from the following herbaria: BM, BR, FHO, K,
MO and P, and from WAG where the study was undertaken (species investigated and
voucher details available on request from B. Mackinder). In addition to the described
species of Talbotiella, Hymenostegia, and the four species awaiting description, species
from five other Detarioid genera (Cynometra, Leonardoxa, Loesenera, Plagiosiphon and
Scorodophloeus) were also surveyed. These five additional genera are considered to have
generic affinities with Hymenostegia and Talbotiella based on morphological similarities
and/or as indicated by a maximum parsimony phylogenetic analysis based on trnL data
(Bruneau et al., 2001).
Most of the vegetative characters surveyed were not informative in the study group,
their ranges of variation being continuous with, overlapping with, or contained within,
the ranges of other taxa in the study. However, six characters (in combination) showed
considerable utility in recognising taxa within the group (Table 2). The pattern of
variation of these six characters among the taxa studied is presented in Table 3.
Talbotiella species share a unique combination of vegetative characters with a core
group of eight Hymenostegia species which includes H. floribunda, the type species of the
genus. From hereon we shall refer to this core group as Hymenostegia sensu stricto (Table 4).
Variation in the stipule base was only seen in Talbotiella gentii, some of which
presented auriculate bases, some lacked these and occasionally both states were seen in
the same collection. Nonetheless, the plasticity of stipules observed in this species does
not preclude a hypothesis that, based on a shared combination of vegetative characters,
Hymenostegia sensu stricto and Talbotiella form a clade, but stipule morphology does not
provide evidence as to whether the genera are congeneric or sister to each other. In
either case the hypothesis is based on the assumption that the shared characters are
synapomorphies. However, were we to sample even more widely, it is possible that the
unique character combination would be discovered in other, more distant taxa. If the
shared characters were synplesiomorphies, they would not support a Hymenostegia sensu
stricto /Talbotiella clade.
The unique combination of characters shared by Hymenostegia sensu stricto and
Talbotiella is not found in the other taxa thus far included which collectively present
seven different character combinations and are collectively hereafter referred to as “ex
Hymenostegia” (Table 4). The considerable diversity of character combinations present
45
05Mackinder
11/2/10
3:54 pm
Page 46
Systematics and Conservation of African Plants
TABLE 2. Vegetative characters used for assessing generic limits in Talbotiella, Hymenostegia,
Cynometra, Leonardoxa, Loesenera, Plagiosiphon and Scorodophloeus. The six characters
which, in combination, were most taxonomically useful in this study are emboldened.
stipule fusion; free or fused basally
stipule insertion; lateral or interpetiolar
stipule auricle; present or absent
gland like appendages; present or absent on stipules or on bud scales
leaflet number; minimum number of leaflets per leaf
leaflet number; maximum number of leaflets per leaf
leaflet glands on underside of distal half of leaflets; number
leaflet glands on underside of proximal half of leaflets; number
hydathodes; present or absent
leaflet indumentum upper; absent or present (describe if present)
lower leaflet surface indumentum; absent or present (describe if present)
leaflet indumentum margin; absent or present (describe if present)
leaflet apex shape; acute, rounded or emarginate
leaflet base proximal shape; rounded or angular
midvein; symmetrical or oblique
midvein ending; becoming obscure towards apex, clearly reaching the apex.
leaf rachis; terete, striate, canaliculate or winged
leaf rachis top indumentum; absent or present (describe if present)
leaf rachis lower surface indumentum; absent or present (describe if present)
leaflets sub-opposite (at least some pairs); present or absent
position along the leaf rachis of the largest pair of leaflets; calculated as percentage of
the total number of leaflet pairs per leaf (juga/total jugae) × 100
leaflet length to width ratio; ratio in longest leaflet
leaflet length to width ratio; ratio in shortest leaflets
in these taxa lead us to conclude that Hymenostegia, as currently circumscribed, is a
heterogeneous assemblage. The heterogeneity of Hymenostegia as traditionally
circumscribed had been previously been noted by Bruneau et al. (2001), who sampled
eight species (five Hymenostegia sensu stricto species and three ex Hymenostegia species)
but only now after wider sampling, is the extent of the full heterogeneity apparent.
TABLE 4. Species assigned to Hymenostegia sensu stricto species and those excluded from
the genus (ex Hymenostegia species) based on vegetative characters.
Hymenostegia sensu stricto species
H. breteleri
H. felicis
H. floribunda
H. klainei
H. mundungu
H. neoaubrevillei
H. normandii
H. pellegrinii
46
ex Hymenostegia species
H. afzelii
H. aubrevillei
H. bakeriana
H. brachyura
H. gracilipes
H. laxiflora
H. ngounyensis
H. talbotii
05Mackinder
Dist.
Prox.
Lower leaflet
surface indumentum
Leaf rachis
shape
Talbotiella batesii Baker f. and T. eketensis Baker f.
free
auriculate
1–3
0
appressed puberulous
canaliculate
Talbotiella gentii Hutch. & Greenway
free
auriculate and
not auriculate
1–3
0
appressed puberulous
canaliculate
Hymenostegia spp. including type (H. floribunda)
but excluding species listed below
free
auriculate
1–7
0 (–2)
appressed puberulous
canaliculate
Hymenostegia aubrevillei Pellegr.
free
not auriculate
2–3
1
appressed puberulous
terete to
narrowly
canaliculate
Hymenostegia afzelii Harms
fused
not auriculate
4–7
1–4
glabrous
winged
Hymenostegia laxiflora Harms
fused
not auriculate
0–5
1–2
glabrous
canaliculate
Hymenostegia bakeriana Hutch. & Dalziel
and Hymenostegia talbotii Baker. f.
fused
not auriculate
0–19
0–3
glabrous
terete
Hymenostegia brachyura (Harms) J. Léonard
free
not auriculate
2–6
0
pubescent
canaliculate
Hymenostegia gracilipes Hutch. & Dalziel
fused
not auriculate
4–12
1
appressed puberulous
canaliculate
Hymenostegia ngounyensis Pellegr.
free
not auriculate
0
0
glabrous
striate
Cynometra sensu stricto (C. cauliflora L.)
fused
not auriculate
0
0
scattered pubescent
not applicable
Cynometra sens. lat. (C. hankii Harms)
free
not auriculate
1–3
0
scattered pubescent
sub-winged
Leonardoxa africana (Baill.) Aubrév.
fused
not auriculate
0
0
scattered puberulous
terete
47
Loesenera spp.
free
not auriculate
1–3
1
appressed puberulous
terete
Plagiosiphon spp.
free
not auriculate
0–2
0 (–1)
glabrous
canaliculate
Scorodophloeus zenkeri (Taub.) J. Léonard
free
not auriculate
1–4
0
appressed puberulous
canaliculate
Page 47
Stipule base
3:54 pm
Stipule
fusion
Clarifying the generic limits of Talbotiella and Hymenostegia
Taxa
11/2/10
TABLE 3. Selected vegetative characters as observed in Talbotiella, Hymenostegia and putatively related taxa. The first three rows share a
character combination which is not found in other taxa in the study. Each of the other rows contains a unique combination of characters.
Dist. = number of distal leaflet glands; Prox. = number of proximal leaflet glands.
05Mackinder
11/2/10
3:54 pm
Page 48
Systematics and Conservation of African Plants
4 Evidence from nucleotide sequence data
Two datasets are available, rbcL and trnL which may be relevant to resolving the generic
limits of Talbotiella and Hymenostegia.
RbcL.
A model-based (Mr Bayes) phylogenetic analysis using rbcL sequences (Lunenburg,
2006) included a wide sampling from within Hymenostegia and Talbotiella as well as from
many other detarioid genera (Fig. 1).
The final alignment of 62 accessions (for voucher details see Appendix 1) and 734
characters (available from Lunenburg) yielded 77 parsimony informative characters.
Alignment was straightforward with no insertions or deletions, but a lack of variation
resulted in few informative characters and concomitantly produced a relatively
unresolved and weakly supported phylogeny. Nonetheless, Lunenburg’s (2006)
analyses recovered a clade comprising all six Hymenostegia sensu stricto species sampled
(Clade 1). Two of the three accessions of H. felicis were in this clade but a third
accession was anomalously placed with Plagiosiphon multijugus and Scorodophloeus zenkeri,
possibly due to lab error or contamination. A second clade (Clade 2) was recovered that
included T. gentii, (the only described Talbotiella successfully amplified) and three of the
species awaiting description (the fourth failed to amplify) suggesting their affinity with
Talbotiella, along with two ex Hymenostegia species (Clade 2).
Lunenburg’s analysis was congruent with the findings of the study based on
vegetative morphology as to which species belong in Hymenostegia sensu stricto and
which do not. Six accessions representing four other ex Hymenostegia species were
placed outside Clades 1 and 2. The two accessions of Hymenostegia ngounyensis were
placed together in a clade that also contained Anthonotha macrophylla P. Beauv. and
Berlinia occidentalis Keay; the positions of Hymenostegia afzelii, H. brachyura and H.
aubrevillea were unresolved.
The relationship of the two weakly supported Clades 1 and 2 were unresolved.
However, Lunenburg (2006) generated a consensus network (not shown) using the
software Splitstree3 that indicated that, in at least a third of the equally likely Bayesian
tree topologies obtained, Clade 1 was not sister to Clade 2.
TrnL.
The maximum parsimony phylogenetic analyses of Bruneau et al. (2001) based on trnL
nucleotide sequence data, were designed to investigate higher level relationships across
subfamily Caesalpinioideae. One supplementary finding of these analyses was that
Hymenostegia, as currently circumscribed, was not recovered as monophyletic. Of the
eight Hymenostegia species sampled by Bruneau et al. (2001), the five Hymenostegia sensu
stricto species, including the type species H. floribunda, were resolved in a clade with T.
gentii (the only Talbotiella available for sampling), Leonardoxa africana and H. talbotii (ex
Hymenostegia) placed as closest relatives although their individual positions were not
resolved. The two further ex Hymenostegia species sampled, H. afzelii and H. ngounyensis
were placed outside this grouping and not together.
48
05Mackinder
11/2/10
3:54 pm
Page 49
Clarifying the generic limits of Talbotiella and Hymenostegia
Clade 1
Clade 2
77
96
98
64
97
63
97
78 57
58
91
62
79
99
83
99
99
53
100
100
67 97
88
92
99
100
95
90
Hymenostegia felicis 1
Hymenostegia normandii 1
Hymenostegia mundungu
Hymenostegia pellegrinii
Hymenostegia normandii 2
Hymenostegia neoaubrevillea
Hymenostegia klainei 1
Hymenostegia klaniei 2
Hymenostegia klaniei 3
Hymenostegia felicis 2
Talbotiella gentii
undescribed sp. nov. 1
Hymenostegia gracilipes ♦
Hymenostegia bakeriana ♦
undescribed sp. nov. 2
undescribed sp. nov. 3-1
undescribed sp. nov. 3-2
Lebruniodendron leptanthum 1
Lebruniodendron leptanthum 2
Hymenostegia cf pellegrini
Neochevalierodendron stephan
Crudia gabonensis
Hymenostegia ngounyensis 1 ♦
Hymenostegia ngounyensis 2 ♦
Anthonotha macrophylla
Berlinia occidentalis
Cynometra mannii
Cynometra iripa
Cynometra letestui
Plagiosiphon multijugus
Hymenostegia felicis 3
Cynometra lujae
Plagiosiphon gabonensis
Scorodophloeus zenkeri 1
Hymenostegia afzelii ♦
Scorodophloeus zenkeri 2
Scorodophloeus fischeri
Cynometra ananta
Plagiosiphon discifer
Hymenostegia brachyura ♦
Plagiosiphon longitubus
Cynometra sp. nov.
Hymenostegia afzelii 2 ♦
Saraca palembanica
Zenkerella citrina
Normaniondendron bequaertii
Afzelia quanzensis
Scorodophloeus zenkeri 3
Loesenera gabonensis
Plagiosiphon sp. nov.
Hymenostegia aubrevillea ♦
Brownea sp. nov.
Macrolobium acaciifolium
Brownea ariza
Brownea coccinea
Tamarindus indica
Polygala amara
Polygala chamaebuxus
Cercis siliquastrum
Bauhinia candicans
Rourea coccinea
Connarus conchocarpus
FIG. 1. Bayes consensus tree from the rbcL analyses reproduced from Lunenburg, 2006. Bayesian
posterior probabilities above 50 are shown above the line and Jackknife support values above 50%
are shown below. Multiple accessions of the same taxon are distinguished by numbers e.g. 1, 2 etc.
following the specific epithet. Taxa considered in this study are emboldened. Ex Hymenostegia
species are marked with black diamonds.
49
05Mackinder
11/2/10
3:54 pm
Page 50
Systematics and Conservation of African Plants
5 Evidence from floral morphology
Léonard’s postulate (1951) that Talbotiella and Hymenostegia are parts of the same
evolutionary series was founded on observations of petal number. In Hymenostegia (as
understood in 1951) he recorded a range of (1) 2–5 petals for the genus. After
undertaking numerous dissections of the described species of Hymenostegia we
conclude that the number of well developed, showy petals is a reliable character for
distinguishing species. Furthermore, with one exception (discussed below),
Hymenostegia sensu stricto species produce either two or three well-developed
spathulate petals.
Of the species of Hymenostegia in which Léonard observed a sequential reduction
in petal number, all but H. gracilipes (ex Hymenostegia) possess two or three welldeveloped spathulate petals, 7–10 × c. 3 mm. Nevertheless, during floral dissections it
was not unusual for the authors to encounter one or two rudimentary petals which
never exceeded 6 × 0.5 mm and were usually shorter and narrower. The presence or
absence of these rudimentary petals (which Léonard included to derive petal
numbers of four and five, including five for the type species H. floribunda) varied
among collections of the same species and even within the same accession. We
therefore suggest the erratic presence of these structures renders them unsuitable for
inclusion in a quantitative evolutionary series. Thus we consider the species of
Hymenostegia sensu stricto to be characterised by the presence of two or three welldeveloped spathulate petals providing a clear distinction from Talbotiella species
which lack petals.
An exception to the pattern of producing two to three well-developed petals is
Hymenostegia breteleri Aubrév., the only member of the genus to be described since
Léonard published his observations in 1951. This rare species was not included in the
trnL analysis (Bruneau et al., 2001), nor was it successfully amplified for the rbcL study,
so there is no molecular data available to test its placement in Hymenostegia sensu stricto.
During the inselberg excursion held during the AETFAT congress this species endemic
to the Yaoundé area was recollected, enabling its incorporation in future studies.
Hymenostegia breteleri possesses the combination of vegetative characters that are
common to the species of Hymenostegia sensu stricto and Talbotiella but apetaly and pollen
characters (see below) suggest that it is misplaced in Hymenostegia and should be
transferred to Talbotiella.
6 Evidence from pollen morphology
Pollen of Hymenostegia afzelii and H. floribunda has been previously illustrated as part of
a survey of Detarieae pollen (Banks & Klitgaard, 2000). For the present study, we
examined pollen from two species of Talbotiella (the type T. eketensis and T. gentii), four
species of Hymenostegia sensu stricto (the type H. floribunda as well as H. klainei, H.
mundungu and H. breteleri, and two of the undescribed species, using a scanning electron
microscope (SEM).
The pollen of both Talbotiella species had narrow apertures and was fully tectate. In
contrast, the species of Hymenostegia sensu stricto (not including H. breteleri) had pollen
with broader apertures and was only semi-tectate. Pollen from the type species of the
two genera is shown for comparison in Figs 2A and 2B. Pollen of two of the undescribed
species (Figs 2C and 2D) resembles the pollen of T. eketensis (the type species of
Talbotiella) as does that of Hymenostegia breteleri supporting the hypothesis that H. breteleri
is more closely related to the species currently placed in Talbotiella.
50
05Mackinder
11/2/10
3:54 pm
Page 51
Clarifying the generic limits of Talbotiella and Hymenostegia
A
B
C
D
E
FIG. 2. Scale bars are 10 microns. A Hymenostegia floribunda. SEM of whole grain in equatorial view
showing broader apertures and semi-tectate striate surface; B Talbotiella eketensis. SEM of whole
grain in oblique polar view showing narrow apertures and tectate striate surface; C Talbotiella sp.
nov. 1. SEM of whole grain in oblique polar view showing narrow apertures and tectate striate
surface; D T. sp. nov. 4. SEM of whole grain, equatorial view also showing narrow apertures and
tectate striate surface; E Hymenostegia breteleri. SEM of whole grains showing narrow apertures and
tectate striate surface.
51
05Mackinder
11/2/10
3:54 pm
Page 52
Systematics and Conservation of African Plants
7 Summary of characters examined in this study
A summary of the morphological and palynological evidence presented in this paper is
given in Table 5. With respect to the characters historically used to distinguish Talbotiella
and Hymenostegia (Table 1) i.e. bud scale persistence, bracteole shape and persistence,
and petal number, the status of Talbotiella is unchanged. However, with the exclusion of
eight species from Hymenostegia sensu lato and the reassessment of petal number, the
more narrowly circumscribed Hymenostegia sensu stricto becomes morphologically more
homogenous for bud scale persistence and petal number.
TABLE 5. Summary of the revised and extended character set relevant to the
determination of the generic limits of Hymenostegia sensu stricto and Talbotiella, including,
for comparison, the character states of the four undescribed species and the misplaced
H. breteleri.
Characters
Hymenostegia
sensu stricto
(excluding
H. breteleri)
Persistent
imbricate
bud scales
Persistent
narrow
petaloid
bracteoles
Hymenostegia
breteleri
Talbotiella
Undescribed
species
always present
sometimes
present
always
present
always
present
sometimes
present
Petals
2–3
0
0
0
Pollen
apertures broad,
ornamentation
semi-tectate
and striate
apertures narrow, ornamentation
tectate and striate
8 Decision to maintain Talbotiella as distinct from
Hymenostegia sensu stricto
The character combination of free auriculate stipules, 1–3 glands on the distal part of
the leaflet, glands absent from the proximal part, appressed puberulous indumentum
on the lower leaflet surface indumentums, and a canaliculate leaf rachis is shared by the
species of Hymenostegia sensu stricto, H. breteleri, Talbotiella spp. and the four undescribed
species. Based on these characters alone, a case could be made for combining Talbotiella
with Hymenostegia sensu stricto and H. breteleri and including the undescribed species
within Hymenostegia. However, given the absence of petals in Talbotiella, Hymenostegia
breteleri and the undescribed species, along with pollen possessing narrow apertures and
a tectate surface, and in light of the results from the molecular studies, we have decided
to uphold Talbotiella as distinct. The undescribed species are all apetalous and pollen
sampled from two of the undescribed species matches that of Talbotiella, as does the
pollen of the apetalous Hymenostegia breteleri.
52
05Mackinder
11/2/10
3:54 pm
Page 53
Clarifying the generic limits of Talbotiella and Hymenostegia
We therefore consider it best to place the undescribed species in Talbotiella and we
propose transferring Hymenostegia breteleri to Talbotiella, thus expanding this genus from
three to eight species.
9 The relationship of Talbotiella to Hymenostegia sensu stricto
Although we have chosen to maintain Hymenostegia sensu stricto and Talbotiella as distinct,
their relationship remains unclear. The shared vegetative characters (listed above)
could be interpreted as indicating a sister group relationship, but due to lack of
resolution from the Bayesian analysis of the rbcL sequence data that interpretation is
neither supported nor refuted (Fig. 1) and the results of the trnL studies of Bruneau et
al. (2001) were no less equivocal. Moreover, without a more extensive morphological
phylogenetic study that includes a larger number of outgroup genera, we can not rule
out the possibility that these shared characters are not synplesiomorphies. An analysis
of Bayesian consensus topologies uncovered evidence that Hymenostegia sensu stricto and
Talbotiella may not be sisters (Lunenburg, 2006), which lends support to maintaining
Talbotiella as distinct. A combined molecular and morphological phylogenetic analysis
of Caesalpinioideae (Herendeen et al., 2003) did not place Talbotiella (represented by T.
gentii) as sister to a clade comprising the five sampled species of Hymenostegia sensu stricto
including the type H. floribunda, but instead placed Hymenostegia afzelii (ex Hymenostegia)
as sister to Hymenostegia sensu stricto species with limited support.
10 Summary of results
Talbotiella can be distinguished from Hymenostegia sensu stricto.
Hymenostegia senso lato, to which 16 species are currently assigned, is a heterogeneous
assemblage.
Hymenostegia sensu stricto comprises seven species, H. floribunda (the type), H. felicis, H.
klainei, H. neoaubrevillei, H. normandii, H. mundungu and H. pellegrinii.
Hymenostegia breteleri is misplaced in Hymenostegia and should be transferred to Talbotiella.
The species misplaced in, and eventually to be excluded from Hymenostegia (ex
Hymenostegia species) are H. afzelii, H. aubrevillei, H. bakeriana, H. brachyura, H. gracilipes,
H. laxiflora, H. ngounyensis and H. talbotii.
The eight ex Hymenostegia species do not appear to comprise a natural group.
11 Future research plans
Based on the results presented here, we plan to pursue the following activities. We will
publish a revision of Talbotiella to include 8 species, three of which are already described
in the genus, four will be newly described and one species will be transferred from
Hymenostegia. The pollen data set will be expanded by undertaking wider sampling and
recording more characters. We will build a widely sampled trnL dataset and add key taxa
to the rbcL dataset, ultimately to be analysed in combination in an effort to resolve to
determine the correct placement of ex Hymenostegia species and to elucidate the
relationship between Talbotiella and Hymenostegia sensu stricto. We plan also to publish a
revision of Hymenostegia sensu stricto.
53
05Mackinder
11/2/10
3:54 pm
Page 54
Systematics and Conservation of African Plants
Acknowledgements
We thank an anonymous reviewer whose careful and constructive comments improved
the manuscript. We would also like to thank the curators of BM, BR, FHO, K, MO and
P for the loan of the relevant sheets, the Head of the National Herbarium of Cameroon
for access to the herbarium collection at YA, and the Curator of the National
Herbarium of Ghana for permitting Yvette Harvey to take digital images of the
specimens at GC. This work was supported by a NL-TAF SYNTHESES grant which
allowed the first author to make a study visit to WAG.
References
Baker, E. G. (1914). Talbotiella gen. nov. J. Bot. 52: 2.
Baker, E. G. (1929). Talbotiella batesii Baker, sp. nov. J. Bot. 68: 197.
Banks, H. & Klitgaard, B. B. (2000). Palynological contribution to the systematics of
detarioid legumes (Leguminosae: Caesalpinioideae). In: P. S. Herendeen & A.
Bruneau (eds), Advances in Legume Systematics 9: 79–106. Royal Botanic Gardens, Kew.
Bruneau, A., Forest, F., Herendeen, P. S., Klitgaard, B. B. & Lewis, G. P. (2001).
Phylogenetic relationships in the Caesalpinioideae (Leguminosae) as inferred from
chloroplast trnL intron sequences. Syst. Bot. 26: 487–514.
Cowan, R. S. & Polhill, R. M. (1981). Detarieae DC. In: R. M. Polhill & P. H. Raven
(eds), Advances in Legume Systematics 1: 117–134. Royal Botanic Gardens, Kew.
Herendeen, P. S., Bruneau, A. & Lewis, G. P. (2003). Phylogenetic relationships in
Caesalpinioid legumes: A preliminary analysis based on morphological and
molecular data. In: B. B. Klitgaard & A. Bruneau (eds), Advances in Legume
Systematics 10: 37–62. Royal Botanic Gardens, Kew.
Hutchinson, J. & Dalziel, J. M. (1928). Flora of West Tropical Africa. H.M.S.O., London.
Léonard, J. (1951a). Hymenostegia. Bull. Jard. Bot. État Bruxelles 21: 433–445.
Léonard, J. (1951b). Talbotiella. Bull. Jard. Bot. État Bruxelles 21: 445–446.
Léonard, J. (1957). Genres des Cynometreae et des Amherstieae africaines (Leguminosae
- Caesalpinioideae). Essai de blastogénie appliqueés de la systématique. Mém. Acad.
Roy. Belgique, Cl. Sci. 30(2): 1–312.
Lunenburg, I. (2006). Molecular phylogenetic analysis of Hymenostegia (Fabaceae;
Caesalpinioideae; Detarieae) and related genera. Possible implications for generic
classification of Hymenostegia species. Biosystematics Group. Wageningen, NHN
Wageningen University Branch. MSc. thesis.
Mackinder, B. A. (2005). Detarieae. In: G. P. Lewis, B. D. Schrire, B. A. Mackinder &
J. M. Lock, Legumes of the World, pp. 69–109. Royal Botanic Gardens, Kew.
54
05Mackinder
11/2/10
3:54 pm
Page 55
Clarifying the generic limits of Talbotiella and Hymenostegia
APPENDIX 1. Voucher details for rbcL sequences adapted from Lunenburg (2006).
GenBank ID is reported for taxa when relevant.
Taxon
Voucher
GenBank ID. or
Herbarium code
Afzelia quanzensis Welw.
Anthonotha macrophylla P. Beauv.
Bauhinia candicans Benth.
Berlinia occidentalis Keay
Brownea ariza Benth.
Brownea coccinea Jacq.
Brownea sp. nov.
Cercis siliquastrum L.
Connarus conchocarpus F. Muell.
Crudia gabonensis Pierre ex De Wild.
Cynometra ananta Hutch. & Dalziel
Cynometra iripa Kostel.
Cynometra letestui (Pellegr.) J. Léonard
Cynometra lujae De Wild.
Cynometra mannii Oliv.
Cynometra sp.
Hymenostegia afzelii (Oliv.) Harms 1
Hymenostegia afzelii (Oliv.) Harms 2
Hymenostegia aubrevillei Pellegr.
Hymenostegia bakeriana Hutch. & Dalziel
Hymenostegia brachyura (Harms)
J. Léonard
Hymenostegia felicis (A.Chev.) J. Léonard 1
Hymenostegia felicis (A.Chev.) J. Léonard 2
Hymenostegia felicis (A.Chev.) J. Léonard 3
Hymenostegia gracilipes Hutch. & Dalziel
Hymenostegia klainei Pierre ex Pellegr. 1
Hymenostegia klainei Pierre ex Pellegr. 2
Hymenostegia klainei Pierre ex Pellegr. 3
Hymenostegia mundungu (Pellegr.)
J. Léonard
Hymenostegia neoaubrevillei J. Léonard
Hymenostegia ngounyensis Pellegr. 1
Hymenostegia ngounyensis Pellegr. 2
Hymenostegia normandii Pellegr. 1
Hymenostegia normandii Pellegr. 2
Hymenostegia pellegrinii (A. Chev.)
J. Léonard
Hymenostegia cf. pellegrinii (A. Chev.)
J. Léonard
Lebruniodendron leptanthum (Harms)
J.Léonard 1
Lebruniodendron leptanthum (Harms)
J. Léonard 2
Goyder 3727
Wieringa 4590
Breteler 13966
van Valkenburg 3006
van Valkenburg 3028
van Andel 4214
Breteler 13352
van Andel 4244
de Kruif 668
van der Burgt 587
Zenker 4481
WAG; YA
WAG; LBV
Z70161
GC; WAG
Z70158
Z70159
U74186
WAG
U06798
LBV; WAG
WAG
AY289677
LBV; WAG
LBV; WAG
LBV; WAG
WAG; YA
WAG
WAG; YA
WAG
WAG; YA
WAG; YA
Wieringa 2581
Wieringa 2841
Jacques-Félix 5129
Merello 1299
Wieringa 4714
Wieringa 2575
Wieringa 5137
Devred 2640
LBV; WAG
LBV; WAG
WAG; YA
GC; WAG
LBV; WAG
LBV; WAG
LBV; WAG
BR; WAG
Breteler 14793
Wieringa 2579
Wieringa 4493
Breteler 13943
Wieringa 2208
Breteler 15602A
LBV; WAG
LBV; WAG
LBV; WAG
LBV; WAG
WAG; YA
LBV; WAG
Wieringa 3771
LBV; WAG
Wieringa 1659
LBV; WAG
Allogho 20
LBV; WAG
Jongkind 2068
Wieringa 5680
Wieringa 2585
Wieringa 4296
55
05Mackinder
11/2/10
3:54 pm
Page 56
Systematics and Conservation of African Plants
Loesenera gabonensis Pellegr.
Macrolobium acaciifolium Benth.
Neochevalierodendron stephanii
(A. Chev.) J. Léonard
Normandiodendron bequaertii (De Wild.)
J. Léonard
Plagiosiphon discifer Harms
Plagiosiphon gabonensis (A. Chev.)
J. Léonard
Plagiosiphon longitubus (Harms)
J. Léonard
Plagiosiphon multijugus (Harms)
J. Léonard
Plagiosiphon sp. nov.
Polygala amara L.
Polygala chamaebuxus L.
Rourea coccinea (Thonn. ex Schumach.)
Benth.
Saraca palembanica (Miq.) Baker
Scorodophloeus fischeri (Taub.)
J. Léonard
Scorodophloeus zenkeri Harms 1
Scorodophloeus zenkeri Harms 2
Scorodophloeus zenkeri Harms 3
Talbotiella gentii Hutch. & Greenway
Talbotiella sp. nov 1
Talbotiella sp. nov 2
Talbotiella sp. nov 3-1
Talbotiella sp. nov 3-2
Tamarindus indica L.
Zenkerella citrina Taub.
56
Wieringa 4542
Breteler 13262
LBV; WAG
U74191
LBV; WAG
LBV; WAG
J. J. F. E. de Wilde
11209
Wieringa 2316
Wieringa 4400
WAG; YA
LBV; WAG
not recorded
WAG
Wieringa 3813
LBV; WAG
Breteler 12828
LBV; WAG
Z70175
Z70176
AF308704
Breteler 13509
Faulkner 4025
WAG
WAG
Breteler 14073
Wieringa 5067
Breteler 11111
Jongkind 1940
Letouzey 10940
van der Burgt s.n.
van der Burgt 601
van der Burgt 708
LBV; WAG
LBV; WAG
LBV; WAG
GC; WAG
K; P; YA
K; WAG; YA
K; WAG; YA
K; WAG; YA
Z70160
K; WAG; YA
Cheek 7614