Content uploaded by Rafael S. Almeida Pérez
Author content
All content in this area was uploaded by Rafael S. Almeida Pérez on Sep 07, 2020
Content may be subject to copyright.
Botanical Journal of the Linnean Society (1998), 128: 291–314. With 6 figures
Article ID: bt980193
A new species of the wild dragon tree, Dracaena
(Dracaenaceae) from Gran Canaria and its
taxonomic and biogeographic implications
AGUEDO MARRERO
1
∗, RAFAEL S. ALMEIDA
2
AND
MANUEL GONZA
´LEZ-MARTI
´N
3
1
Jardı
´n Bota
´nico Viera y Clavijo, Apartado 14 de Tafira Alta, 35017,
Las Palmas de Gran Canaria, Canary Islands, Spain
2
Seccio
´n de Geografı
´a (DACT), Universidad de Las Palmas de Gran Canaria,
Edificio Humanidades, C/. Pe
´rez del Toro Num. 1, 35003, Las Palmas de Gran Canaria,
Canary Islands, Spain
3
Servicio de Planificacio
´n de Recursos Naturales, Viceconsejerı
´a de Medio Ambiente,
Tafira Baja, 35017, Las Palmas de Gran Canaria, Canary Islands, Spain
Received January 1998; accepted for publication April 1998
The arborescent taxa of Dracaena which form the dragon tree group comprise five species
found in Macaronesia, Morocco (D. draco), East Africa (D. ombet,D. schizantha), Arabia (D.
serrulata) and the island of Socotra (D. cinnabari). A new species of dragon tree, Dracaena
tamaranae A. Marrero, R.S. Almeida & M. Gonza´lez-Martı´n, is described from Gran
Canaria, Canary Islands. This new species differs from D. draco, the only other Dracaena
species currently known in Macaronesia, in having a growth form and inflorescence type
and leaves more similar to the East African and Arabian species of Dracaena. In contrast, D.
draco appears to be related to D. cinnabari. In this paper, we also present a study of the
taxonomy, habitat and ecology of all the species of the dragon tree group. These are found
in thermo-sclerophyllous plant communities of tropical–subtropical regions which are rather
xerophilous and have a rainfall range of 200–500 mm. Our study indicates two independent
colonization events for Dracaena in Macaronesia. In addition, we suggest that the dragon tree
group provides an example of two major biogeographical disjunctions between East and
West Africa. We postulate that this group has a Tethyan origin, a hypothesis supported by
fossil and palaeoclimatic data, and thus parallels the distribution and dispersal pattern of
other taxonomic groups.
1998 The Linnean Society of London
ADDITIONAL KEY WORDS:—Canary Islands – corology – ecology – Macaronesia.
CONTENTS
Introduction ....................... 292
Material and taxonomic references ............... 294
Description ....................... 294
Dracaena tamaranae A. Marrero, R.S. Almeida & M. Gonza
´lez-Martı
´n,
sp. nov. ..................... 294
∗Correspondence to A. Marrero.
291
0024–4074/98/110291+24 $30.00/0 1998 The Linnean Society of London
A. MARRERO ET AL.292
Conservation status .................... 297
Taxonomic discussion .................... 300
Habitat and ecology .................... 303
Biogeographic relationships .................. 306
Dracaena and the Rand Flora ................ 306
The fossil data .................... 307
Panbiogeographic interpretation ............... 308
Acknowledgements .................... 310
References ....................... 310
Appendix ........................ 313
INTRODUCTION
The genus Dracaena comprises approximately 60 species (Mabberley, 1990)—50
species sensu Friis (1992)—which are mainly found in tropical and subtropical Africa.
At least 23 species occur in the Guinea-Congo region in western Africa (Bos, 1984).
The genus also reaches Macaronesia, Arabia, Socotra, Madagascar, southeastern
Asia, northern Australia, and one species (D. americana Donn. Sm.) is found in the
neotropics.
The dragon tree group is formed of five arborescent species (i.e. D. cinnabari Balf.
f., D. draco (L.) L., D. ombet Kotschy & Peyr., D. serrulata Baker and D. schizantha
Baker), and provides one of the best known examples of disjunct distribution between
Macaronesia, Morocco and East Africa.
Until the discovery of the new species described here, Dracaena draco was the only
species found in Macaronesia. It occurs in the Madeira archipelago, where it has
been reported for the islands of Madeira and Porto Santo, although it is currently
regarded as extinct in the latter. This species also thrives in the Canary Islands,
where it is currently found on all the islands. However, wild populations of this
species are only known in Tenerife and Gran Canaria; it is likely, therefore, that
the present populations of D. draco in the rest of the archipelago are of cultivated
origin. Dracaena draco also reaches the Cape Verde Islands where it is found on the
islands of Sa
˜o Nicolau, Santo Anta
˜o and Fogo, and it is considered extinct on the
islands of Sa
˜o Vicente and Sa
˜o Tiago (Bystro
¨m, 1960). Dracaena draco has been
recently discovered in southern Morocco, in the Anti-Atlas region, where it is
regarded as a distinct taxon: D. draco subsp. ajgal Benabid et Cuzin (Benabid &
Cuzin, 1997).
Dracaena draco has been considered as closely related to the three species found in
Socotra, D. cinnabari and East Africa, D. ombet (D. orabet sensu Engler, 1908), and D.
schizantha (Balfour, 1883; Christ, 1885; Baker, 1898; Sunding, 1970). It is noteworthy
that there is another species in Arabia (D. serrulata) which is usually ignored in
accounts of the genus for the Red Sea region. These four East African and Arabian
species have an allopatric distribution. Dracaena cinnabari is endemic to Socotra island
where it thrives in the northeastern mountain range of Haggier (Balfour, 1883),
mainly in the highlands of Mumi (Beyhl, 1995). Dracaena serrulata has a scattered
distribution along the southwestern edge of the Arabian Peninsula, mainly in the
hills of southern Medina and the El Asir mountains in Saudi Arabia; in the foothills
of the highlands of Yemen and on northern slopes of Dhofar at Oman (Collenette,
1985; Miller & Cope, 1996). Dracaena ombet grows along the African hills which face
the Red Sea; it is found in Jebel Elba in southeastern Egypt, Mount Erkowit in
Sudan, escarpments of the Eritrean mountains and in the mountains of Djibouti
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 293
Figure 1. Distribution of the seven living species and three fossil species of the dragon tree group.
(Baker, 1897, 1898; Ta
¨ckholm & Drar, 1973; Friis, 1992). Lastly, D. schizantha is
found in north-facing escarpments of Harar in Ethiopia, in the mountains of Djibouti
and in the northern mountains of Somalia to the Ahl Mountains where it almost
reaches the Horn of Africa (Baker, 1877; Friis, 1992; Thulin, 1995). A further
species, D. hanningtoni Baker, restricted to Mozambique, also appears to be related
to the five species which comprise the dragon tree group (Baker, 1898) (Fig. 1).
Wild dragon trees are extremely rare in Gran Canaria, and it was only in the
early 1970s that they were reported for the first time, in the western mountains of
Tauro and in the Barranco de Arguineguı
´n (Kunkel, 1972, 1973). Most recently
the distribution of this species has been mapped by Rodrigo & Montelongo (1986).
All these references identified the individuals as D. draco.
In the last 5 years one of us (R.S. Almeida) has undertaken extensive field studies
relating to the distribution, phytogeography and conservation status of the wild
dragon trees on Gran Canaria. As a result of these studies we were able to collect
ripe fruits in 1994. Seedlings cultivated from these studies led us to the belief that
the Gran Canaria individuals could not, in general, belong to D. draco.
A. MARRERO ET AL.294
There are several morphological differences between the seedlings of these two
species. The seedling leaves of D. draco are flat whereas those of the new taxon are
conduplicate. In addition, the young root of the Gran Canaria endemic is napiform.
These obvious differences in juvenile characters did also extend to other traits of
adults such as the shape and habit of the inflorescence or the articulation of the
pedicels.
MATERIAL AND TAXONOMIC REFERENCES
To conduct this study we have used morphological studies of living material and
also of herbarium specimens (Appendix). It is also based on a critical review of
several morphological descriptions of the different species of the dragon tree group:
Linnaeus, 1767; Webb & Berthelot, 1836–50; Hooker & Smith, 1851; Schweinfurth,
1868; Baker, 1877, 1894, 1897, 1898; Bentham & Hooker, 1880; Balfour, 1883,
1888; Engler, 1908; Brown, 1914; Andrews, 1956; Maire, 1958; Thonner, 1962;
Ta
¨ckholm & Drar, 1973; Beyhl, 1995; Thulin, 1995; Turland, 1995; Benabid &
Cuzin, 1997.
DESCRIPTION
Dracaena tamaranae A. Marrero, R.S. Almeida & M. Gonza
´lez-Martı
´n, sp.
nov.
arborescens robusta, 6–10 m alta, ramificatione primaria trichotoma, rare
tetrachotoma et ramificationibus posterioribus dichotomis vel simplicibus.
flavo-griseus, vix signis foliaribus notatus, leviter nitidus. foliis equitantibus
bilateralibus compressis, radicibus valde succulentis, primariis cylindrico-globosis,
secundariis napiformibus. subulata, canaliculata, leviter falciformia, 40–80
(110)×3–4.5 cm, glauca, subtus leviter striata, margine omnino hyalina, ad basim
incrassata cum pseudovagina fusco-rubinea, subamplexicauli, atque arcu manifesto et
angusto, 10 (8–11) cm longo. paniculata complexa, glabra, tripinnata,
gracilis, 80–100 cm, per totam longitudinem ramificata. brevissimus.
basales foliis similes, ad apicem cito decrescentes in formam diminutam,
primum ensiforme demum subulatae et lineares, acuminatae. 2–5 in fasciculis
dispositi. 2.25–3.25 mm longi, ad apicem articulati. minutae,
triangulares vel ovato-triangulares. 9–15 mm, laete albo-viride; tepala
oblongo-linearia, interna paulum angustiora quam exteriora, basi connati tubo
brevissimo. quam tepala breviora, ad stigma sub anthesi adiacentia; filamenta
6.5–9 mm, connata ad 2 mm, leviter medio incrassata, non complanata; antherae
2 mm, flavo-virides. 3-loculare, 3.6×2.4 mm, ovulo in quoque loculo
solitario. stipitatum. filiformis quam ovarium longior, 5.8 mm,
stigmate capitato trilobulato. glauco-virides, demum aurantiaci, globosi,
10–11.5 mm, vulgo monospermi. globosa vel late ovoidea, leviter compressa,
6–7 mm.
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 295
Figure 2. Holotype of Dracaena tamaranae A. Marrero, R.S. Almeida & M. Gonza
´lez-Martı
´n, sp.
nov.
Type. Dracaena tamaranae A. Marrero, R.S. Almeida & M. Gonza
´lez-Martı
´n. Habitat
in Canaria Magna (Gran Canaria dicta)in loco dicto “barranquillo Andre
´s”, 825 m supra
mare,loc. class. leg.: A. Marrero, M. Gonza
´lez-Martı
´n & A. Quintana, die 31.vii.1997
(LPA:18525, holotypus in MA, Fig. 2). Isotypi: ibidem (duplicata in LPA, TFC, K), idem,
R.S. Almeida, A. Marrero & A. Quintana, 20.vii.1997 (LPA: 18524 cum duplicata in
MA, ORT, BM). (Icon: Fig. 3).
Additional material examined. Dracaena tamaranae A. Marrero, R.S. Almeida & M.
Gonza
´lez-Martı
´n, Gran Canaria, San Bartolome
´de Tirajana, los Vicentillos, leg.
A. MARRERO ET AL.296
Figure 3. Icon:Dracaena tamaranae sp. nov. A, habit. B, seedling. C, leaves. D, terminal branch
with inflorescence. E, flowers. F, bracts. G, bracteoles. H, tepals. I, pistil. J, stamens. K, fruits. L,
seeds.
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 297
R.S. Almeida & A. Marrero, 17.vii.1997 (LPA: 18521); Ibidem, Moga
´n, barranco de
Arguineguı
´n, Los Pen
˜ones, leg. R.S. Almeida & A. Marrero, 3.vii.1997 (LPA: 18517);
Ibidem, los Gavilanes, R.S. Almeida, A. Marrero & M. Gonza
´lez-Martı
´n, 11.vii.1997
(LPA: 18518, 18519); Ibidem, El Palmarete, R.S. Almeida, A. Marrero & M.
Gonza´lez-Martı´n, 11.vii.1997 (LPA: 18520); Ibidem, Barranquillo Andre´s, leg. R.S.
Almeida, A. Marrero & A. Quintana, 20.vii.1997 (LPA: 18522); Ibidem,ex horto,
193–194/96, A. Marrero, 26.vi.1997 (LPA: 18526, 18527) (seedling); Ibidem, barranco
de Moga
´n, Los Laerones, leg. R.S. Almeida, A. Marrero & M. Gonza
´lez-Martı
´n,
19.viii.1997 (LPA: 18523).
arborescent, robust, 6–10 m high, primary branching trichotomous, rarely
tetrachotomous and subsequent branches dichotomous or simple. yellow-grey,
barely marked by foliar scars, slightly glossy. : leaves equitants, bilateral
and compressed, roots highly succulent with a cylindrical globose primary root and
napiform secondary roots (Fig. 4A). 40–80 (110) cm long, 3–4.5 cm wide,
subulate and canaliculate, rather falcate, glaucous, rather striate below, hyaline-
white entire margin, swollen at the base with a basal brown-reddish pseudo-sheath
which is subamplexicaul and forms a patent and narrow arc 10 (8–11) cm long.
80–100 cm long, panicle complex, glabrous, tripinnate, slender,
branches dispersed along the main axis (Fig. 4B). very short. basals
are similar to the mature leaves, secondary bracts decrease rapidly in size, ensiform
to subulate and linear, acuminate. 2–5 clustered. 2.25–3.25 mm
long, articulate towards the apex. minute, triangular or ovate-triangular.
9.5–11 mm, bright greenish-white, tepals oblong-linear, inner tepals nar-
rower than outer ones, joined at the base, tube very short. shorter than
tepals and adjacent to the stigma during anthesis; filaments 6.5–9 mm, joined 2 mm
from the base, slightly swollen in the middle, unflattened; anthers 2 mm, yellow-
greenish. trilocular 3.6×2.4 mm, with a single, stipitate ovule per locule.
5.8 mm, filiform, longer than the ovary; stigma capitate, trilobulate.
10–11.5 mm globose, greenish, glaucous but orange when ripe, usually mono-
spermous. 6–7 mm globose to broadly ovoid and slightly compressed (Tables
1–4).
Dracaena tamaranae is a species restricted to the island of Gran Canaria, Canary
Islands. This species is found between 400 and 900 m altitude, and along an arc in
the southwestern region, from Fataga valley in the southern slopes to La Aldea
valley in the west.
CONSERVATION STATUS
This species is extremely rare and is known from few localities. We propose that
it merits CR status (Critically Endangered) in the IUCN (IUCN, Red List Categories,
1994). Although some populations are located within the network of Gran Canaria
nature reserves (Act 12/94 Espacios Naturales de Canarias) there are still many
unprotected individuals. There is a strong need to establish a rescue programme to
preserve the genetic integrity of this species.
A. MARRERO ET AL.298
Figure 4. A, seedling of Dracaena tamaranae sp. nov. (left) and D. draco (right). B, inflorescence
of D. tamaranae sp. nov.
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 299
T 1. Qualitative data for growth form and leaves of species of the dragon tree group
Basal portion
General appearance Leaves of the leaves
Length Width
along from
Species Aspect Branching Shape Colour Margin Blade Venation the base the base
D. tamaranae robust developed slightly subulate glaucous entire hialine slightly succulent, patent very long narrow
trunk dense greenish to canaliculate and rigid
greyish grooved
D. draco very robust developed dense ensiform green entire reddish slightly succulent, flat, patent long wide
trunk glaucous slightly flexible
D. draco subsp. very robust, very dense ensiform glaucous entire slightly succulent, flat, patent long wide
ajgal developed trunk greenish hardly flexible
D. cinnabari very robust developed very dense slightly subulate to green serrulate towards very succulent, flat, not patent very long wide
trunk linear ensiform glaucous the base and rigid with a thick and
hialine sharp apex
D. ombet robust short trunk slightly ensiform linear to glaucous serrulate hialine slightly succulent, not patent short very wide
dense linear subulate greenish slightly canaliculate,
yellowish semi-rigid
D. schizantha robust short trunk slightly linear subulate glaucous serrulate slightly succulent, not patent long very wide
dense greenish slightly canaliculate,
greyish rigid
D. serrulata robust short trunk slightly ensiform linear to glaucous serrulate slightly succulent, not patent long wide
dense linear subulate greenish slightly canaliculate,
semi-rigid
A. MARRERO ET AL.300
T 2. Qualitative data for flower traits of species of the dragon tree group
Inflorescence Pedicels
Species Ramification Hairiness Bracts Bracteoles Articulation n
o
/cluster
D. tamaranae tripinnate glabrous subulate-linear, triangular-ovate, apical 3–5
along the axes acuminate minute extreme
D. draco bipinnate glabrous linear-subulate, subulate ovate, middle 4–7
below the acuminate acuminate
middle
D. draco subsp. bipinnate glabrous linear-subulate, subulate ovate, middle 4–7
ajgal below the acuminate acuminate
middle
D. cinnabari bipinnate glabrous subulate to triangular-ovate, middle 3–7
middle subulate- acuminate extreme
acuminate
D. ombet tripinnate glabrescens triangular, ovate minute middle –
along the axes acuminate
D. schizantha tripinnate felted subulate triangular, middle to 1–4
along the axes tomentose triangular, acuminate middle
acuminate, minute extreme
D. serrulata tripinnate tomentose or subulate, triangular, middle to 1–5
along the axes dense triangular, subulate middle
tomentose acuminate acuminate extreme
Perianth Stamens Fruits
Species Segment Tube Colour Anthers Filament Colour
D. tamaranae oblong linear very short whitish green
1
3
–
1
4
slightly orange
filament thickened, not
flattened
D. draco subspatulate short whitish pink
1
2
–
1
3
thickened orange
linear campanulate filament flattened reddish
D. draco subsp. subspatulate short whitish yellow
1
2
–
1
3
thickened red
ajgal linear campanulate filament flattened orange
D. cinnabari oblong basal and pale greenish twice subulate red
barely yellow filament scarlet
developed
D. ombet linear very short, whitish pink subequal – –
barely filament
developed
D. schizantha lanceolate very short, white
1
2
–
1
3
subulate –
barely filament thickened
developed
D. serrulata – – – – – orange
TAXONOMIC DISCUSSION
Dracaena tamaranae sp. nov. seems to be closely related to the three species found
in the Horn of Africa and Arabia (i.e. D. ombet,D. schizantha and D. serrulata). All
these species have glaucous leaves, minute bracteoles and are not densely branched.
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 301
T 3. Quantitative data for growth form, leaves and inflorescences of species of the dragon tree
group. l/w=length/width
Leaves
Species Height Length Width l/w Base
(m) (cm) (cm) (cm)
D. tamaranae 6–10 40–80 (110) 3–4.5 (5) 10–18 9.8–19.4×3.7–5.5
D. draco 8–12 40–90 (110) 2–3.5 (4) 20–25 6–10.4×3–4.2
D. draco subsp. ajgal 10–20 60 3 20 –
D. cinnabari 6–10 30–60 2.5–4.5 12–13.3 6.5–13.5×2.2–5
D ombet 2–8 35–65 2.2–3 16–21.7 4.5–8×3.2–5.6
D. schizantha 2–9 35–70 0.7–2.5 28–50 4.5×2.4–3.7
D. serrulata 2–8 30–60 2–3.5 15–17 8–10×2.5–4.5
Inflorescence
Species Length Flower pedicels Fruit pedicels Bracteoles
(cm) (mm) (mm) (mm)
D. tamaranae 80–100 2.25–3.25 – 1.90
D. draco 60–120 4.00–6.00 (10) 8.30–10.50 0.80–2.80
D. draco subsp. ajgal – 1.00–4.00 – –
D. cinnabari 30–75 3.00–7.00 6.00–8.00 1.00–4.00
D ombet 20–40 3.10–4.50 semipedicel 0.50–1.50
2.50–4.50
D. schizantha 45 1.50–3.00 3.50–5.00 0.50–1.50
D. serrulata – – 2.50–5.00 0.50–1.50
T 4. Quantitative data for flower and fruit traits of species of the dragon tree group
Flowers Stamens
Ovary Fruit Seed
Species Perianth Tube Filament Anther length diameter size
(mm) (mm) (mm) (mm) (mm) (mm) (mm)
D. tamaranae 9.50–11.00 0.75–1.00 6.08–8.75 1.75–2.00 3.50–3.75 10.00–11.50 6.00–7.00
D. draco 7.00–9.50 1.30–3.50 5.80–8.00 1.00–1.50 3.60–4.50 14.30–14.60 7.50–10.00
D. draco subsp. ajgal 7.00–8.00 1.00–2.00 – – – – 7.50–10.00?
D. cinnabari 5.00–9.50 – – 2.00–2.25 3.50–4.00 8.00–13.00 3.50–5.00
D. ombet 6.30–6.40 – – – – – –
D. schizantha 3.50–6.50 – 3.20–4.75 1.00–1.50 2.75–3.00 5.00–7.00 –
D. serrulata 5.00 – – – – 5.00–8.00 3.50–5.50
In addition, the tripinnate, slender and erect inflorescences are branched along all
the axes.
These continental species have flowers and inflorescences which are smaller than
those of D. tamaranae. Furthermore, they also have rather succulent and linear-
ensiform leaves with acutely serrulated margin and no patent nerves, the trunks are
usually shorter and the pedicels are articulated in the middle section. Moreover, D.
schizantha tends to have linear leaves which are extremely narrow and greyish in
colour, it has tomentose panicles which are dense and short. The basal pseudo-
sheaths of the leaves of D. ombet are only slightly developed, and this species has
A. MARRERO ET AL.302
glabrescent inflorescences. Plants of D. serrulata are distinguished from the Gran
Canaria species by their tomentose inflorescence.
There are several morphological characters which differentiate D. tamaranae from
D. draco and D. cinnabari. These two latter species have ensiform leaves which are
flat and not as glaucous, their inflorescences are bipinnate and robust. In addition,
the branching of the inflorescences is basal. Other unique features of these species
are pedicels articulated in the middle, smaller perianth, flattened filaments and a
robust and dense growth-form. Furthermore, D. cinnabari has extremely rigid leaves
with finely serrate margin towards the base, and nerves which are not patent. In
contrast D. draco usually has reddish leaf-margins, inflorescence branches which tend
to be reflexed or patent after fruiting, linear, acuminate inflorescence bracts and a
whitish-pink perianth with a longer tube. Dracaena hanningtoni seems to be closely
related to D. ombet (Baker, 1898), and it is rather different from the other species of
the dragon tree group in that it has a much longer perianth, with the segments
twice as long as the tube, the stamens are as long as the segments and the style is
exserted. However, the general shape of the inflorescence indicates that this species
seems to be closely related to the D. draco–D. cinnabari group.
We consider that D. tamaranae has strong morphological relationships with the
East African and Arabian species. The main morphological features which support
this hypothesis are the type of inflorescence and the general habit. However, we
postulate that D. draco is closely related to the Socotra species, D. cinnabari. This
assumption would mean that there are two biogeographical disjunctions between
Macaronesia and the Red Sea region. A phylogenetic confirmation of this hypothesis
would provide one of the few cases for independent colonizations of oceanic islands
by congeneric species. Previous studies based on molecular data have proven this
to be the case for Lavatera (Malvaceae) in the Canary Islands (Ray, 1995; Fuertes-
Aguilar et al., 1996) and for Gossypium (Malvaceae) in Galapagos (Wendel & Percival,
1990; Wendel & Percy, 1990). It is likely that other genera such as Convolvulus,
Euphorbia,Limonium,Salvia,Senecio or Viola, with endemic species in Macaronesia,
have also colonized these islands more than once.
Two species, D. draco and D. tamaranae, occur in western Africa. The former occurs
both on the mainland and the Macaronesian Islands, whilst the latter is restricted
to Gran Canaria. The occurrence of the insular species can only be explained
through long-distance dispersal because these islands appear to be oceanic and
therefore have never been joined to the continent (Aran
˜a & Carracedo, 1978; Aran
˜a
& Ortiz, 1984; Carracedo, 1984). It is likely that these two species remained in a refuge
in the Macaronesian Islands, and following a decrease of continental populations of
dragon trees from western Africa. D. draco is now only found in a very restricted
area of southern Morocco. It is noteworthy that these Morocco populations have
been given subspecies rank as D. draco subsp. ajgal, an indication that the taxon is
in early stages of speciation (Fig. 5A).
The geological events which led to the formation of the African Great Rift have
had an impact in the current distribution of dragon trees in East Africa. In this
region the Great Rift splits into two major fissures, one towards the Red Sea, the
other towards the Aden Gulf. These geological events began in the early Miocene
and resulted in the fragmentation of the original populations of dragon trees. These
new sub-populations probably initiated new speciation events which yielded the
three species which currently thrive in Arabia (D. serrulata), the African hills of the
Red Sea (D. ombet) and the Horn of Africa (D. schizantha). These three species are
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 303
AFRICA
Azores
Km
0 1000 020
20
40
D. tamaranae
D. draco
D. draco
Madeira
Canary Islands
Macaronesia
Cape Verde
Islands
D. draco
A
AFRICA
Km
0 1000
D. ombet
D. serrulata
Socotra
B
ARABIA
D. schizantha
50
D. cinnabari
10
30
Red Sea
Gulf of Aden
30
Figure 5. A, dispersal and colonization from the African mainland make the Macaronesian Islands a
refuge for several species of Dracaena. B, African Grand Rift leads to the disruption and vicariance of
Dracaena in the Red Sea region.
closely related, and they might form a species complex; indeed, some authors have
suggested that they are conspecific with D. ombet s.l. (Deil, 1988; Friis, 1992; Thulin,
1995). However, Miller & Cope (1996) considered D. serrulata to be a distinct species.
Our preliminary studies indicate that D. schizantha also has some morphological
differences. In contrast, D. cinnabari is a distinct species which has several unique
morphological features. It is restricted to the continental island of Socotra. This
island was part of the Horn of Africa before the Middle Pliocene (Fig. 5B).
Saporta (1862, 1865, 1873a,b) described three Neogene fossil species of dragon
tree (Dracaenites brongniartii Saporta, D. sepultus Saporta and D. narbonensis Saporta)
from southern France, which he suggested were closely related to D. draco. However,
at that time the dragon trees of East Africa and Arabia were not known to the
scientific community. Dracaenites brongniartii, like Dracaena draco and D. cinnabari, has
ensiform leaves. Plants of D. ombet s.l. have linear-ensiform leaves, but they are never
truly ensiform. The two other fossil species have strictly linear leaves and, therefore,
they may be related to those species which have subulate-linear or linear-ensiform
leaves (Table 5). All the species have foliar scars in the bark, and these are particularly
evident in the D. ombet group and D. cinnabari. The foliar scars have a rough margin
in D. cinnabari. In contrast, they are tenuous in the fossil species and in D. tamaranae.
Dracaenites sepultus has an extremely warty bark, a feature which is unknown in any
of the living species. However, we are aware that based on these morphological
features, it is difficult to establish phylogenetic links between the fossil species and
the present ones. There is no reason to assume that these fossils represent the direct
ancestors of the living species. Sadly there are no fossils of fruits and flowers which
would help to clarify this issue.
HABITAT AND ECOLOGY
Dracaena tamaranae sp. nov. is found in the thermophile zone where it grows on
inaccessible slopes and cliffs which tend to be shady and humid. The open Juniperus
A. MARRERO ET AL.304
T 5. A selection of the most important morphological features of several fossil species (Dracaenites)
from the French Tertiary
Leaf branches
Fossil taxon Locality Stipe Diameter Bark Foliar scars
D. narbonensis Armissan arboreal 10 cm slightly rough slightly patent
without protuberance
D. sepultus Aix, Provence bush — warty with very tenuous
protuberance
D. brongniartii Aix, Provence giant 10 cm — —
D. minor Aix, Provence frutescens — — very patent
Leaves
Fossil taxon Shape Outward Venation Length Width Base
appearance
D. narbonensis linear flat with margin delicately nerved 150 cm 4 cm gradually expanded
entire 10–12 cm
D. sepultus linear firm nerved — — broad
D. brongniartii ensiform rigid and firm nerved striated — 3 cm clearly broad
10–12 cm
D. minor strictly shiny and firm very tenuous extremely — little and sharply
linear delicate long expanded
compressed
unequal
bushland (Oleo-Rhamnetalia crenulatae Santos 1983) and the Cistus scrubland (Cisto-
Micromerietalia P. Pe
´rez et al. 1991) form a mosaic in this area. Species which grow
in this zone are: Juniperus turbinata Guss. subsp. canariensis (Guyot) Rivas Mart.,
Wildpret & P. Pe
´rez, Olea europaea L. subsp. cerasiformis (Webb & Berthel.) G. Kunkel
& Sunding, Teline rosmarinifolia Webb & Berthel., Globularia cf. salicina Lam., etc.
(Rodrigo & Montelongo, 1986; Marrero, Gonza
´lez-Artiles & Gonza
´lez-Martı
´n,
1995). Species which are characteristic of northern and northeastern humid slopes
of the archipelago also occur in this area. Among these are: Davallia canariensis (L.)
J.E. Sm., Pericallis webbii (Sch. Bip.) Bolle, Sonchus acaulis Dum. Cours., Hypericum
canariense L., Ranunculus cortusifolius Willd., etc. Dracaena tamaranae reaches the dry
canary pine forest (Cytiso-Pinetea canariensis Rivas Goday & Esteve ex Sunding 1972)
at higher altitude. This species can also be found at lower altitude among elements
of the Kleinio-Euphorbietalia canariensis (Rivas Goday & Esteve 1965) Santos 1976. The
average annual rainfall of the zones where D. tamaranae grows is 200–350 mm.
Both in Macaronesia and Morocco, D. draco lives in areas which are not as xeric
as those of D. tamaranae. For instance, on Madeira, it is mainly found as an element
of thermo-sclerophyllous zones between sea-level and 200 m, and on sea-facing cliffs
where some species of the laurel forest also occur (Turland, 1995). On Tenerife, D.
draco also grows in thermo-sclerophyllous zones (Mayteno-Juniperion canariensis Santos
& Ferna
´ndez Galva
´nex Santos 1983) between 100 and 600 m. Other species found
in this zone are: Juniperus turbinata subsp. canariensis,Maytenus canariensis (Loes.) G.
Kunkel & Sunding, Rhamnus crenulata Aiton, Olea europaea subsp. cerasiformis,Pistacia
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 305
atlantica Desf., Globularia salicina, etc. (Santos, 1983; Rivas Martı
´ez et al., 1993). In
addition, Sideroxylon marmulano Banks ex Lowe and Apollonias barbujana (Cav.) Bornm.,
two thermophile species of the laurel forest, are also found in this zone. The annual
average rainfall of this zone ranges between 200 and 400 mm. On Gran Canaria
this species grows in similar areas, but it is extremely rare. On Cape Verde Islands,
Dracaena draco is found between 700 and 1000 m where it grows together with
Euphorbia tuckeyana Steud. ex Webb, Echium spp., Ficus sur Forssk. (F. capensis Thunb.),
F. sycomorus L. subsp. gnaphalocarpa (Miq.) C.C. Berg, Sideroxylon marmulano,Acacia
albida Del., etc. (Bystro
¨m, 1960). The Anti-Atlas populations of D. draco are located
between 400 and 1300 m with Laurus azorica (Seub.) Franco, Davallia canariensis,
Rhamnus alaternus L., Quercus rotundifolia Lam., Ceratonia siliqua L., Teline segonnei (Maire)
Reynaud, Argania spinosa Skeels, Olea maroccana Greuter & Burdet, etc. These species
are part of the association Davallio canariensis–Dracaenetum ajgal Benabid & Cuzin 1997
in the order Acacio-Arganietalia Barbe
´ro et al. 1982 (Benabid & Cuzin, 1997). Annual
rainfall is around 400 mm. However, Rivas-Goday & Esteve-Chueca (1965) defined
the thermo-sclerophyllous shrubs with Dracaena draco as the climax of the Crassi-
Euphorbietea,Diacanthio-Euphorbietea (Kleinio-Euphorbietea).
The ecology of these species is similar to that of all the East African and Arabian
species. For example, D. ombet is usually found in mountain escarpments together
with O. europaea subsp. africana (Mill.) P.S. Green (including O. chrysophylla Lam.),
Euclea racemosa Murr., Euphorbia abyssinica J.F. Gmel., Acacia etbaica Schweinf., A. tortilis
(Forssk.) Hayne, Ziziphus spina-christi (L.) Desf. and Lycium arabicum Schweinf. (Kassas,
1956; White, 1983). Dracaena ombet occurs in plant communities of Dracaeno-Eu-
phorbietalia abyssinicae Knapp 1968 (Deil & Mu
¨ller-Hohenstein, 1984) which are
situated below the evergreen scrub zone where there are thermo-sclerophyllous
species such as Maytenus senegalensis (Lam.) Exell and Euclea schimperi (DC.) Dandu,
and widespread species such as Euphorbia abyssinica,Acacia etbaica and A. tortilis (Kassas,
1956). Annual average rainfall of these zones is 200 mm.
Dracaena schizantha grows on escarpments along the northern mountains of Somalia.
It is found in transition plant communities, which are situated between afromontane
forests and evergreen sclerophyllous scrubs. In these transition plant communities
the following species are found: Olea europaea subsp. africana (including O. chrysophylla
and O. somaliensis Baker), Juniperus procera Hochst. ex Endl., Acokanthera schimperi Oliver,
Pistacia aethiopica Kokwaro, (P. lentiscus L. subsp. emarginata Engl.), P. falcata Mart.,
Osyris lanceolata Hochst. & Steud. ex DC., Euphorbia abyssinica,Monotheca buxifolia DC.
(Sideroxylon gillettii Hutch. & E.A. Bruce), Maytenus undata (Tunb.) Blakelock, etc. (Fici,
1991; Friis, 1992). Annual average rainfall reaches 500 mm.
Dracaena serrulata is a very rare species which is found in the xerophile zone, in
an area where Acacia–Commiphora bushland is dominant. This zone is situated either
below the deciduous forest of Acacia or forms a mosaic with bushes of Olea europaea
subsp. africana and Juniperus procera. This dragon tree occurs in the Red Sea and
Aden Gulf mountains of the Arabian Peninsula. Populations are located on sea-
facing slopes and on inner regions of these mountains. In the xerophile zone are
found several species of Acacia and Commiphora:A. niotica (L.) Willd. ex Del., A. etbaica,
A. gerrardii Benth., A. tortilis, etc.; C. habessinica (Berg) Engl., C. foliaceae Sprague, C.
gileadensis C. Chr., etc. Other species include Maerua crassifolia Forssk., Ziziphus spina-
christi,Euphorbia balsamifera Aiton subsp. adenensis (Delf) Bally, E. cuneata Vahl, E.
triaculeata Forssk., etc. (Miller & Cope, 1996). Average annual rainfall is approximately
200 mm.
A. MARRERO ET AL.306
Dracaena cinnabari is found on slopes of the highlands of northeastern Socotra.
This area is mainly covered by thickets of Rhus thyrsiflora Balf. f., Cephalocroton socotranus
Balf. and Allophyllus rhoidiphyllus Balf. f. Other species present are Boswellia ameero
Balf. f., B. socotrana Balf. f., Jatropha unicostata Balf. f. and Croton socotranus Balf. f. At
higher altitude, this type of thicket is intermixed with Hypericum shrubland. All the
plant communities are under the influence of the northeastern humid monsoons
(White, 1983; Beyhl, 1995; Miller & Cope, 1996). Mean average rainfall is ap-
proximately 400 mm.
Bystro
¨m (1960) suggested that all the dragon tree species share similar ecological
requirements. They tend to grow in areas with average temperatures of 18–20°C.
They are found between 10°N in Somalia and 33°N in Madeira, and there is a
clear correlation between latitude and altitude. The populations of Madeira may
be found at sea-level whilst those of Somalia never occur below 1400–1800 m.
In general, all these arborescent species with an umbrella-shaped canopy are
found mainly on the margins of the tropical-subtropical regions. They are part of
a thermo-sclerophyllous vegetation similar to the Canarian Oleo-Rhamnetalia crenulatae
or to the Arabian communities of Acacia–Commiphora. These plant communities are
usually intermixed with xerophilous formations which are similar to the Canarian
Kleinio-Euphorbietalia. They are mostly linked to steep and rocky landscapes, but there
are some ecological differences between them.
Dracaena tamaranae grows in more xeric and hotter areas than D. draco. Populations
of the latter tend to be affected by the northeastern trade winds and also grow
under more humid conditions. The most xeric species of warm environments are
D. serrulata and D. ombet. In contrast, D. cinnabari is the most mesophilic species. It
grows along a belt of the highlands of Socotra which has the highest levels of rainfall.
This habitat has similar features to those found in the region where D. draco is
located in Macaronesia and Morocco.
BIOGEOGRAPHIC RELATIONSHIPS
Balfour (1883) and Christ (1885) have previously suggested close taxonomic
relationships between the members of the dragon tree group. Other authors use
this group as one of the best examples of biogeographic disjunction between
Macaronesia and East Africa (Meusel, 1965; Sunding, 1970, 1979; Bramwell, 1986).
Hooker (1878) was the first to propose that the dragon tree, together with other
species of the Macaronesian laurel forest, are relicts of an old vegetation which once
existed in northwest Africa. Axelrod (1975) is in agreement with this idea, and
proposed that subtropical elements, such as Dracaena and Sideroxylon, found refuge in
East and West Africa, as a consequence of the desertification of the Sahara in the
late Oligocene.
Dracaena and the Rand Flora
The flora which existed in the southwestern region of South Africa during the
Palaeocene was defined by Lebrun as the Rand Flora (cf. Que
´zel, 1978; Que
´zel &
Barbe
´ro, 1993). According to Que
´zel (1978, 1983) some of the elements which
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 307
currently thrive in northern Africa might be considered part of this ancient flora.
These elements include species of the following genera: Aristida,Amphinomia,Andrachne,
Gaillonia,Periploca,Tribulus,Trichodesma,Zygophyllum,Asthenaterum,Oropetium,Enneapogon,
Coccullus,Neurada,Ifloga,Oligomeris, etc. Some of the xerophilous taxa which currently
exist in Macaronesia, the Red Sea region and the Saharan highlands have been
assigned to this flora (Que
´zel, 1978, 1983). These xerophilous taxa include Euphorbia
balsamifera, and several species of genera such as: Euphorbia of the cactiform types,
Acacia,Ceropegia,Commelina,Dracaena,Kalanchoe,Kleinia,Mesembryanthemum,Pentzia,
Wahlenbergia, etc.
Que
´zel (1978, 1983) reviewed the most important biogeographical components
of the North African region, and he considered D. draco as one of the elements of
the Rand Flora. However, Que
´zel (1978) proposed that the East African D. ombet
is linked to another floristic group associated with “tropical elements of mountain
massifs next to coast”. In addition, most recently Bramwell (1986, 1990) included
all the species of the dragon tree group among the elements of the Rand Flora.
Que
´zel (1978) suggested that many of the elements of the Rand Flora reached
the Sahara region using the East Africa mountains as a dispersal route during the
Oligo-Miocene, and this is the reason why there are several present disjunctions
between northern and southern Africa. During this process there was an intermixing
with northern elements. It is well known that there have been many climatic
fluctuations in northern Africa since the late Oligocene, and these major climatic
changes had a drastic effect on the vegetation of this region (Aubre
´ville, 1970, 1976;
Axelrod, 1973; Que
´zel & Barbe
´ro, 1993), and, therefore, the fact that the flora of
this region has several components should be borne in mind. Some of these
components originated on the shores of the African Tethys which had a tropical
flora similar to that found today in the Congo-Guinea region. Other North African
elements are related to the laurasian subtropical flora, whereas some groups come
from the dry tropical flora of the Sahara (Axelrod, 1975; Aubre
´ville, 1976). There
is no doubt that some of the North African species are linked to the Rand Flora;
however, we believe that the most primitive species of the dragon tree group may
have been derived from any of these flora. Fossil evidence could provide new insights
into this issue.
The fossil data
None of the Tertiary Sahara fossils found so far belong to any of the Rand Flora
groups (Maley, 1980). However, there are at least six species of Dracaena (Dracaenites)
from remains of the French Eocene and Neogene (Saporta, 1862, 1865, 1873a,b,
1888, 1889). Three of them are considered as members of the dragon tree group
(i.e. D. brongniartii,D. narbonensis and D. sepultus). One of the other three fossil species,
D. resurgens Saporta, has been proposed to be closely related to the shrubby D.
angustifolia Roxb. (Saporta, 1889), a species of the xerophytic group (Engler, 1908;
Mies, 1995). It seems feasible from the morphology of the two remaining fossil
species (D. minor Saporta and D. pusillus Saporta) that they may be also associated
with the xerophytic group. However, these two fossil species are dwarf and have
very long, narrow leaves.
Two additional extinct Dracaena species have been identified based on pollen from
the Neogene (Van Campo & Sivak, 1976). Pollen of the first species, D. saportae Van
A. MARRERO ET AL.308
Campo & Sivak comes from Bohemia whilst the second one, D. guinetii Van Campo
& Sivak, was found in Tunez. These two species seem to be related to those which
are currently found in the Guinea–Congo region. Dracaena saportae appears to be
related to Dracaena ovata Ker Gawl. whilst D. guinetii is associated with D. humilis
Baker (Van Campo & Sivak, 1976).
From these palaeobotanical data it is obvious that the only fossils which seem
referable to the dragon tree group come from the Eocene–Neogene European
Tethyan area. These fossils have been found together with other elements of the
subtropical laurasian forest which existed in southern Europe during the Tertiary.
These are related to some of the taxa which currently thrive in the Macaronesian
laurel forest (Saporta, 1862, 1865, 1873b, 1889; Depape, 1922; Andrea
´nszky, 1968;
Takhtajan, 1969; Sunding, 1970, 1979; Bramwell, 1972, 1976; Axelrod, 1975). The
only fossil available from the African Tethyan is clearly related to Dracaena taxa
which currently grow in tropical Africa. We are aware that the Dracaena fossil data
do represent a limited sample of all the taxa which existed in the past. However,
palaeobotanical data suggest that at least seven Dracaena species existed in European
Tethyan and also indicate that the genus Dracaena had an important centre of
diversity in this region.
Panbiogeographic interpretation
A biogeographic interpretation of the patterns of distribution of the dragon tree
group can be made using the panbiogeographic approach (Croizat, 1958, 1968).
According to this methodology, and using both data from living and fossil species,
we can establish a route which links Macaronesia with East Africa and Arabia
through the European Tethyan (Fig. 6). It is worth mentioning that Croizat
(1968) previously found a similar route for Sedum sect. Afrosdeum and sect. Epeteium
(Crassulaceae), which runs from Macaronesia–northwest Africa to Mesopotamia,
Ethiopia and Kenya. A similar situation might be found with Aeonium (Crassulaceae:
Sempervivoideae) which occurs in Macaronesia, northwestern Africa and in the
Red Sea region. This genus does not have any species or fossil data from the
Mediterranean region, and it has been proposed as one of the Rand Flora elements
by Bramwell (1986). However, an inclusive study of Aeonium and Sempervivum (Sem-
pervivoideae) supports that both genera have a Tethyan subtropical origin (Meusel,
1965). This is in agreement with recent molecular data which suggest that the tribe
Sempervivoideae is nested within the tribe Sedoideae and Sedum (Hart, 1991; van
Ham, 1995; Mes, 1995; Stevens, 1995). Bearing in mind the hypothesis of an holartic
origin for Sedum (Croizat, 1968; Hart & Eggli, 1995), it seems likely that Aeonium
would also have an origin in this region. This hypothesis had already been proposed
by Fici (1991) who considered that this genus migrated towards southern latitudes
during the Quaternary cool periods. Other examples which seem to follow similar
routes are Lavandula and Coris monspeliensis L. (Fici, 1991). The data presented in this
paper indicate that Dracaena and Aeonium may have similar dispersal routes (cf.
Meusel, 1965). These two genera seem to have an origin in a thermo-sclerophyllous
flora which existed in the Tethyan, and they do not seem to be associated with the
South African eremitic-xerophile Rand Flora.
Our study gives support to the hypothesis that the current species of the dragon
tree group are a depleted and relict representation of the Mio-Pliocene Saharan
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 309
AFRICA
Azores
Km
0 2000
20
20
Madeira
Canary
Islands
Cape Verde
Islands
D. draco–D. cinnabari–D. hanningtoni
1000
FRANCE
MOROCCO
EGYPT
SUDAN
ETHIOPIA
SOMALIA
Socotra
ARABIA
0
20
40
20
Indian Ocean
MOZAMBIQUE
Atlantic Ocean
D. tamaranae–D. ombet "complex"
Origin and dispersal node
040 60
Figure 6. The two routes which link the distribution of the living and fossil species
of the dragon tree group.
xerophile-sclerophyllous flora. These species would have their origin in the most
thermic elements of the Oligo-Miocene laurasian subtropical flora. They existed in
the edges of the forests and it is likely that they occurred on sunny and exposed
areas of rocky slopes, cliffs and escarpments. Following the major climatic changes
of the Miocene, they might have migrated southward where they established new
populations in northern Africa. Further climatic changes might have led to the
gradual disjunction of these populations towards the eastern and western margins
of Africa and towards the Saharan islands mountains.
Aubre
´ville (1976) cites Bombax as a good example of a laurasian group which
could have migrated towards Africa in the Tertiary. This genus with eight species
is currently restricted to tropical regions of Asia (Indo-Malaysia) and Africa. There
is strong palaeobotanical evidence for the existence of Bombax in the Sahara region
(Bombacoxylon) and Europe (Aubre
´ville, 1970, 1976; Saporta, 1862, 1873b). This
seems to indicate that the current African species come from an original Saharan
pool which previously had a Tethyan European laurasian origin (Aubre
´ville, 1976).
Many of the biogeographical studies of Macaronesia and North Africa have
suggested that most of the xerophilous elements of these regions originated from
the Rand Flora (Que
´zel, 1978, 1983; Maley, 1980; Bramwell, 1986, 1990; Que
´zel
A. MARRERO ET AL.310
& Barbe
´ro, 1993). This view has meant that there has been a trend to postulate
strong biogeographic relationships between North Africa–Macaronesia and an
ancient flora which originated in southern Africa in the Palaeocene. This idea has
led workers to underestimate the contribution of the thermo-sclerophyllous Tethyan
flora. It might well be that Dracaena provides an example of this underestimation,
which could also apply to other plant groups.
ACKNOWLEDGEMENTS
We are grateful to Trinidad Arcos (Departamento de Filologı
´a Espan
˜ola, Cla
´sica
y Arabe, Universidad de Las Palmas de Gran Canaria) for her invaluable help with
the Latin diagnosis. We would also like to express our gratitude to Jorge Naranjo
(Viceconsejerı
´a de Medio Ambiente del Gobierno de Canarias) for his help in
translating the German references. J. Francisco-Ortega and D. Bramwell critically
read the manuscript and provided many constructive comments. Logistic help
reaching populations in the wild was provided by Antonio Quintana. This work
could not have been undertaken without the help of S. Owens, P. Wilkin and J.
Lowley (Royal Botanic Gardens, Kew) and the staff(particularly S. Blackmore, R.
Vickery and S. Knapp) of the libraries and herbarium of the Natural History
Museum of London. The Cabildo Insular de Gran Canaria provided financial
support for travel to England to complete this study.
REFERENCES
Andrea
´nszky G. 1968. Neue und interessante tertia
¨re Pflanzenarten aus Ungarn, VI. Acta Botanica
Academiae Scientiarum Hungaricae 14(3–4):219–242.
Andrews FW. 1956. The Flowering Plants of the Sudan. Vol. III (Compositae–Gramineae). Arbroath: T.
Buncle & Co.
Aran
˜a V, Carracedo JC. 1978. Los volcanes de las islas Canarias I: Tenerife. Madrid: Rueda.
Aran
˜a V, Ortiz R. 1984. Volcanologıa. Madrid: CSIC & Rueda.
Aubre
´ville A. 1970. La flore tropicale tertiaire du Sahara. Adansonia, ser. 2, 10(1):9–14.
Aubre
´ville A. 1976. Centres tertiaires d’origine, radiations et migrations des flores angiospermiques
tropicales. Adansonia, ser. 2, 16(3):297–354.
Axelrod DI. 1973. History of the Mediterranean Ecosystem in California. In: Di Castri F, Mooney
HA, eds. Mediterranean Type Ecosystems, Origin and Structure. New York: Springer-Verlag, 225–277.
Axelrod DI. 1975. Evolution and biogeography of Madrean–Tethyan sclerophyll vegetation. Annals
of the Missouri Botanical Garden 62: 280–334.
Baker JG. 1877. Dracaena schizantha Baker sp. nov. In: Trimen H, The Journal of Botany, new series, vol.
VI. London: Ranken & Co., 71.
Baker JG. 1894. Botany of the Hadramaut expedition. Bulletin of Miscellaneous Information, Royal Botanical
Garden, Kew: 342.
Baker JG. 1897. Dracaena ombet Kotschy et Peir. In: Thiselton-Dyer WT, ed. Hooker’s Icones Plantarum,
ser. 4, vol. VI, part. II. London: Dulay & Co., plate 2539.
Baker JG. 1898. Liliaceae: Dracaena. In: Thiselton-Dyer WT, ed. Flora of Tropical Africa, vol. II,
Hydrocharideae to Liliaceae. Ashford, Kent: L. Reeve & Co., 436–450.
Balfour B. 1883. The Dragon’s Blood Tree of Socotra (Dracaena cinnabari Balf. fil.). Transactions of the
Royal Society of Edinburgh 30: 619–623.
Balfour B. 1888. Botany of Socotra. Transactions of the Royal Society of Edinburgh 31: 1–446.
Benabid A, Cuzin F. 1997. Populations de dragonnier (Dracaena draco L. subsp. ajgal Benabid et
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 311
Cuzin) au Maroc: valeurs taxonomique, bioge
´ographique et phytosociologique. Comptes Rendus de
l’Acade
´mie des Sciences, Paris, Sciences de la vie 320: 267–277.
Bentham G, Hooker JD. 1880. Genera Plantarum, III (1). London: Reeve & Co., Williams & Norgate.
Beyhl FE. 1995. Der Drachenbaum und seine Verwandtschaft: II. Der Echte Drachenbaum, Dracaena
cinnabari, von der Insel Sokotra. Der Palmengarten 59(2):140–145.
Bos JJ. 1984. Dracaena in West Africa. Agricultural University Wageningen Papers 84(1):v+1–126.
Bramwell D. 1972. Endemism in the Flora of the Canary Islands. In: Valentine DH, ed. Taxonomy,
Phytogeography and Evolution. London: Academic Press, 141–159.
Bramwell D. 1976. The endemic flora of the Canary Islands. In: Kunkel G, ed. Biogeography and
Ecology in the Canary Islands. The Hague: Dr W Junk bv Publishers, 207–240.
Bramwell D. 1986. Contribucio
´n a la biogeografı
´a de las islas Canarias. Bota
´nica Macarone
´sica 14:
3–34.
Bramwell D. 1990. Panbiogeography of the Canary Islands flora. In: International Symposium on
Biogeographical Aspects of Insularity. Roma: Accademia Nazionale dei Lincei, 157–166.
Brown NE. 1914. Notes on the genera Cordyline,Dracaena,Pleomele,Sansevieria and Taetsia.Bulletin of
Miscellaneous Information, Royal Botanic Gardens, Kew: 273–279.
Bystro
¨m K. 1960. Dracaena draco L. in the Cape Verde Islands. Acta Horti-Gotoburgensis 23: 179–214.
Carracedo JC. 1984. Origen de las islas. In: Concepcio
´n A, ed. Geografı´a de Canarias. Santa Cruz de
Tenerife: Editorial Interinsular Canaria, 55–64.
Christ DH. 1885. Vegetation und Flora der Canarischen Inseln. Botanische Jahrbu
¨cher 6: 458–526.
Collenette S. 1985. An Illustrated Guide to the Flowers of Saudi Arabia. London: Scorpion Publishing.
Croizat L. 1958. Panbiogeography or an Introductory Synthesis of Zoogeography, Phytogeography, and Geology, vols
I–II. Caracas: L Croizat.
Croizat L. 1968. Introduction raisonne´e a` la Bioge´ographie de l’Afrique. Coimbra: Memo
´rias da Sociedade
Broteriana, XX.
Deil U. 1988. Prima¨re und sekunda¨ re Standorte sukkulentenreicher Pflanzengesellschaften in Su
¨dwest-
Arabien. Flora 180: 41–57.
Deil U, Mu
¨ller-Hohenstein K. 1984. Fragmenta Phytosociologica Arabiae-Felicis I. Eine Euphorbia
balsamifera Gesellschaft aus dem jemenitischen Hochland und ihre Beziehungen zu makaronesischen
Pflazengesellschaften. Flora 175: 407–426.
Depage G. 1922. Recherches sur la flore Pliocene de la Valle
´eduRho
ˆne. Annales des Sciences Naturelles,
Botanique 10(4): 73–265.
Engler A. 1908. Die Pflanzenwelt Afrikas. In: Engler A, Drude O, eds. Die Vegetation der Erde, 9(2).
Leipzig: Verlag von Wilhelm Engelmann.
Fici S. 1991. Floristic relations between eastern Africa and the Mediterranean region with special
references to northern Somalia. Flora Mediterranea 1: 175–185.
Friis I. 1992. Forests and Forest Trees of Northeast Tropical Africa. New Bulletin Addition Series XV.
London: HMSO, Royal Botanic Gardens, Kew.
Fuertes-Aguilar J, Ray MF, Francisco-Ortega J, Jansen RK. 1996. Systematics and evolutions
of the Macaronesian endemic Malvaceae based on morphological and molecular evidence. In:
Abstracts II Symposium “Fauna and Flora of the Atlantic Islands”. Las Palmas de Gran Canaria, 51.
Ham RCHJ van. 1995. Phylogenetic relationships in the Crassulaceae inferred from chloroplast
DNA variation. In: Hart H, Eggli U, eds. Evolution and Systematics of the Crassulaceae. Leiden: Backhuys
Publishers, 16–29.
Hart H. 1991. Evolution and classification of the European Sedum species (Crassulaceae). Flora
Mediterranea 1: 31–61.
Hart H, Eggli U. 1995. Introduction: Evolution of Crassulaceae systematics. In: Hart H, Eggli U,
eds. Evolution and Systematics of the Crassulaceae. Leiden: Backhuys Publishers, 7–15.
Hooker JD. 1878. On the Canarian Flora as compared with the Maroccan. In: Hooker JD, Ball J,
eds. Journal of a Tour in Marocco and the Great Atlas. London: Macmillan and Co., 404–421.
Hooker WJ, Smith J. 1851. Dracaena draco. Dragon’s-blood Tree. Botanical Magazine 77: tab. 4571.
IUCN Species Survival Commission. 1994. IUCN Red List Categories. Gland: IUCN.
Kassas M. 1956. The mist Oasis of Erkwit, Sudan. Journal of Ecology 49: 180–194.
Kunkel G. 1972. Sobre el hallazgo del drago (Dracaena draco) en Gran Canaria. Cuadernos de Bota´ nica
Canaria 14–15: 39–42.
Kunkel G. 1973. Novedades en la flora de las islas orientales, con una nota sobre Dracaena draco en
Gran Canaria. Cuadernos de Bota
´nica Canaria 17: 35–37.
A. MARRERO ET AL.312
Linnaeus C. 1767. Systema Natura, ed. 12, 2. Holmiae: Impensis Direct, 736 pp.
Mabberley DJ. 1990. The Plant-Book. A portable dictionary of the higher plants. Cambridge: Cambridge
University Press.
Maire R. 1958. Flore de l’Afrique du Nord, vol. 5: Monocotyledonae: Liliales: Liliaceae. Paris: Paul Lechevalier.
Maley J. 1980. Les changements climatiques de la fin du Tertiaire en Afrique: leur conse
´quence sur
l’apparition du Sahara et de sa ve
´ge
´tation. In: Williams MAJ, Faure H, eds. The Sahara and the Nile.
Quaternary environments and prehistoric occupation in northern Africa. Rotterdam: AA Balkema, 63–86.
Marrero A, Gonza´lez-Artiles F, Gonza´lez-Martı´n M. 1995. Corologı
´a de varias especies raras
de las bandas del sur de Gran Canaria, islas Canarias. In: Notas corolo´ gico-taxono´micas de la
Flora Macarone
´sica (n
os
35–81). Bota
´nica Macarone
´sica 33: 49–63.
Mes THM. 1995. Phylogenetic and systematic implications of chloroplast and nuclear spacer sequence
variation in the Macaronesian Sempervivoideae and related Sedoideae. In: Hart H, Eggli U, eds.
Evolution and Systematics of the Crassulaceae. Leiden: Backhuys Publishers, 30–44.
Meusel H. 1965. Die Reliktvegetation der Kanarischen Inseln in ihren Beziehungen zur su
¨d- und
mitteleuropa
¨ischen Flora. In: Gersch M, ed. Gesammelte Vortra
¨ge u
¨ber moderne Probleme der Abstammunslehre,
1. Jena: Friedrich-Schiller-Universita
¨t, 117–136.
Mies B. 1995. On the comparison of the flora and vegetation of the island groups of Socotra and
Macaronesia. Boletim do Museu Municipal do Funchal (Histo´ria Natural), sup. 4(B): 455–471.
Miller AG, Cope TA. 1996. Flore of the Arabian Peninsula and Socotra, vol. 1. Edinburgh: Edinburgh
University Press, Royal Botanic Garden, Edinburgh and Royal Botanic Gardens, Kew.
Que
´zel P. 1978. Analysis of the flora of Mediterranean and Saharan Africa. Annals of the Missouri
Botanical Garden 65: 479–534.
Que
´zel P. 1983. Flore et ve
´ge
´tation actuelles de l’Africa du nord, leur signification en fonction de
l’origine, de l’e
´volution et des migrations des flores et structures de ve
´ge
´tation passe
´es. Bothalia
14(3–4):411–416.
Que´ zel P, Barbe´ro M. 1993. Variations climatiques au Sahara et en Afrique se
`che depuis le Plioce
`ne:
enseignements de la flore et de la ve´ge´tation actuelles. Bulletin d’Ecologie 24(2–3–4):191–202.
Ray MF. 1995. Systematics of Lavatera and Malva (Malvaceae, Malveae) – a new perspective. Plant
Systematics and Evolution 198: 29–53.
Rivas-Goday S, Esteve-Chueca F. 1965. Ensayo fitosociolo
´gico de la Crassi-Euphorbietea macaronesica
y estudio de los tabaibales y cardonales de Gran Canaria. Anales del Instituto Bota
´nico AJ Cavanilles
22: 221–339.
Rivas-Martı
´nez S, Wildpret W, Del Arco M, Rodrı
´guez O, Pe
´rez de Paz PL, Garcia-Gallo
A, Acebes JR, Dı
´az TE, Ferna
´ndez-Gonza
´lez F. 1993. Las comunidades vegetales de la isla
de Tenerife (islas Canarias). Itinera Geobota
´nica 7: 169–374.
Rodrigo Pe
´rez J, Montelongo Parada V. 1986. Distribucio
´n de especies significativas para la
comprensio
´n de las formaciones boscosas en Gran Canaria (Islas Canarias). I. Bota
´nica Macarone
´sica
12–13: 3–16.
Santos A. 1983. Vegetacio
´n y flora de La Palma. Santa Cruz de Tenerife: Interinsular Canaria.
Saporta MG. 1862. E
´tudes sur la ve´ge´tation du sud-est de la France a l’e´poque tertiaire. III. Flore
des lignites infe
´rieurs, ou e
´tage a
`lignite proprement dit. Annales des Sciences Naturelles, Botanique
4(17):191–311+14 lam.
Saporta MG. 1865. E
´tudes sur la ve
´ge
´tation du sud-est de la France a l’e
´poque tertiaire. 2
a
partie,
III, Flore d’Armissan et de Peyriac, dans le bassin de Narbonne (Aude). Annales des Sciences Naturelles,
Botanique 5(4):5–264+13 lam.
Saporta MG. 1873a. E
´tudes sur la ve
´ge
´tation du sud-est de la France a l’e
´poque tertiaire. Supple
´ment
I, Re
´vision de la flore des gypses d’Aix. II Description des espe
`ces. Annales des Sciences Naturalles,
Botanique 5(17):5–44.
Saporta MG. 1873b. E
´tudes sur la ve
´ge
´tation du sud-est de la France a l’e
´poque tertiaire. Re
´vision
de la flore des gypses d’Aix. Supple
´ment I. Dicotyledoneae. Annales des Sciences Naturelles, Botanique
5(18):23–146+18 lam.
Saporta MG. 1888. Dernie`res adjonctions a la Flore fossile d’Aix-en-Provence. Description des
espe
`ces–1
a
partie. Annales des Sciences Naturelles, Botanique 7(7):1–104.
Saporta MG. 1889. Dernie
`res adjonctions a la Flore fossile d’Aix-en-Provence. Description des
espe
`ces–2
a
partie. Annales des Sciences Naturelles, Botanique 7(10):1–192+30 lam.
Schweinfurth G. 1868. Vegetationsskizzen aus dem su
¨dnubischen Ku
¨stengebirge. Botanische Zeitung
50: 849–860.
NEW SPECIES OF DRACAENA FROM GRAN CANARIA 313
Stevens JF. 1995. Chemotaxonomy of the Eurasian Sedoideae and Sempervivoideae. In: Hart H,
Eggli U, eds. Evolution and Systematics of the Crassulaceae. Leiden: Backhuys Publishers, 45–75.
Sunding P. 1970. Elementer i Kanariøyenes flora, og teorier til forklaring av floraens opprinnelse.
Blyttia 28: 229–256.
Sunding P. 1979. Origins of the Macaronesian flora. In: Bramwell D, ed. Plants and Islands. London:
Academic Press, 13–40.
Ta
¨ckholm V, Drar M. 1973. Flora of Egypt (2nd edn), vol. III: Angiospermae, part Monocotyledones:
Liliaceae–Musaceae. Koenigstein: Otto Koeltz Antiquariat.
Takhtajan A. 1969. Flowering Plants. Origin and dispersal. Edinburgh: Oliver & Boyd.
Thonner FR. 1962. The flowering plants of Africa. In: Cramer J, Swann HK, eds. Historiae Naturalis
Classica. XXVII. New York: Wheldon & Wesley, Ltd and Hafner Publishing Co.
Thulin M. 1995. Dracaenaceae. In: Thulin M, ed. Flora of Somalia, vol. 4: 146. Kew: Royal Botanic
Grdens, 27–30, plate 1-E.
Turland NJ. 1995. Agavaceae. In: Press JR, Short MJ, eds. Flora of Madeira. London: HMSO, Natural
History Museum, 391–392.
Van Campo E, Sivak J. 1976. Pre
´sence de pollens de Dracaena dans le Ne
´oge
`ne me
´diterrane
´en.
Revue de Micropale
´ontologie 18(4):264–268.
Webb PB, Berthelot S. 1836–50. Histoire Naturelle des Iles Canaries. III Botanique. 2 Phytographia Canariensis.
Paris.
Wendel JF, Percival AE. 1990. Molecular divergence in the Galapagos Islands – Baja California
species pair, Gossypium klotzschianum and G. davidsonii (Malvaceae). Plant Systematics and Evolutions 171:
99–115.
Wendel JF, Percy RG. 1990. Allozyme diversity and introgression in the Galapagos Islands endemic
Gossypium darwinii and its relationship to continental G. barbadense.Biochemical Systematics and Ecology
18(7/8): 517–528.
White F. 1983. The Vegetation of Africa. A descriptive memoir to accompany the Unesco/AETFA/UNSO vegetation
map of Africa. Switzerland: Unesco.
APPENDIX
Plant material: Herbarium specimens
Dracaena draco (L.) L., Tenerife, Buenavista del Norte, Ravelo, leg. A. Marrero, R.S. Almeida & M.
Gonza´lez-Martı´n, 31.vii.1997 (LPA: 18504); Idem, Tenerife, Adeje, barranco del Infierno, leg. R.S.
Almeida & A. Marrero, 11.viii.1997 (LPA: 18505); Idem, Gran Canaria, Santa Brı
´gida, barranco
Alonso, Pino Santo (sub-spontaneus), leg. A. Marrero & R.S. Almeida, 4.viii.1997 (LPA: 18503)
(inflorescence, basal bracts, leaves); Idem, Gran Canaria, Jardı´n Bota´ nico Viera y Clavijo (ex horto), leg.
A. Marrero, 22.vii.1997 (LPA: 18506, 18507) (inflorescence, basal bracts, leaves); Idem,(ex horto),
from Cape Verde, leg. A. Marrero, 4.vii.1997 (LPA: 18510, 18511, 18512) (seedling, inflorescence,
infructescence, basal bracts, leaves); Idem, Gran Canaria, El Bata´n–Guiniguada (ex horto), leg. A.
Marrero, 25.vii.1997 (LPA: 18508) (inflorescence, basal bracts, leaves). D. cinnabari Balf. f., Socotra,
Feb.–March, 1880, comm. Prof. Bayley Balfour, Aug. 1880 (K TYPUS) (2 sheets, inflorescence, basal
bracts, leaves, +Ic. Prof. Balfour 5/93); Idem. (BM TYPUS Dupl. inflorescence, basal bracts, leaves);
Idem, Socotra, Dr. Balfour, April, 1880 (K) (infructescence); Idem, Ins. Socotra (Hort. Kew.), coll.
Wikeham Perry, s/n, 19.ix.1878 (K) (leaves); Idem, Socotra, Jebel Shihali, Hagghiher Mts., 3500 ft.,
leg. A.R. Smith & J. Lavranos, 448, 20.iv.1967 (K) (inflorescence); Idem, Socotra, Taukak village, above
Hasen, 450–480 m, leg. M. Thulin & N. Gifri, 8603, 19.i.1994 (K) (infructescence, basal bracts); Idem,
Gran Canaria, Jardı
´n Bota
´nico Viera y Clavijo (ex horto), leg. A. Marrero, 1.ix.1997 (LPA: 18513,
18514) (leaves). D. ombet Kotschy et Peyr., Mount Erkowit, near Suakin, Schweinfurth, 250, 16.ix.1868
(K) (inflorescence, leaves+Hooker’s Ic. Pl. t. 2539); Idem, Gran Canaria, Jardı
´n Bota
´nico Viera y
Clavijo (ex horto), leg. A. Marrero, 22.viii.1997 (LPA: 18515); Idem, 1.ix.1997 (LPA:1‘8516) (leaves); D.
cf. ombet, Djibouti, Wadi Dounyar, S of Ali Sabreh, crest of limestone ridge, 2400 ft., I.S. Collenette,
8644, 20.iv.1993 (K) (inflorescence, leaves). D. serrulata Baker, found on the hills near Dobaibah,
elevation about 4000 ft., coll. W. Lunt., 206, 26.ii.1894 (K TYPUS) (leaves); Idem, Jebel Minmar,
Khawlaan as Sham, c. 2500 m, J.R.I. Wood, Y/75/624, 29.viii.1975 (BM) (2 sheets, infructescence,
leaves); Idem, on the south side of Jebel Minmar (Sa
`dah-Sagayn), c. 2600 m, J.R.I. Wood, 624,
A. MARRERO ET AL.314
29.viii.1975 (K) (infructescence); Idem, Saudi Arabia, the Asir, about 10 km south of Abha, I.S.
Collenette, 628, 6.iv.1978 (K) (leaves); Idem, Saudi Arabia, South Hijaz, Jebel Aba Hassan, a sandstone
massif about 50 km south of the escarpment between Abha-Najran, 5500 ft., I.S. Collenette, 1291,
6.iv.1979 (K) (inflorescence); Idem, Saudi Arabia, S-SW of Madinah, 80 km, 5000 ft. (c. 4600 m), I.S.
Collenette, 3789, 15.viii.1982 (K) (infructescence); Idem, Oman, N of Jabel Qaars, road to Sarfay &
Dhofar, R.M. Lawton, 2398, 28.viii.1982 (K) (2 sheets, infructescence, basal bracts, leaves); Idem,
Duplic. (BM) (infructescence, basal bracts); Idem, Oman, Dhofar, Jebel Semhan above Mirbat, 1350 m,
A.G. Miller & J.A. Nyberg, M-9167, 7.ix.1989 (K) (infructescence, leaves). D. schizantha Baker, Somali-
Land, Meid, Ahl-n. Serrusgeb, 800–1800 m, J.M. Hildebrandt, 1742, April 1875 (BM TYPUS)
(inflorescence); Idem,(KDuplic.) (inflorescence); Idem, Ethiopia, Harar Prov., Steep slopes below Dangago,
15 km SE of Diredawa, along the road to Harar, 1700 m, W. Burger, 1516–1516a, 24.ii.1962 (K) (3
sheets, inflorescence, leaves); Idem, Ethiopia, Harar Prov., Steep slopes below Dangago, 15 km SE of
Diredawa, along the road to Harar, W. Burger, 3714, 4.iii.1965 (K) (2 sheets, inflorescence, leaves);
Idem, Somalia, Valley sides, site A/5 Limestone Mountains, 1340 m, J.B. Billett & R.M. Watson,
23462, 16 & 18.vi.1981 (K) (2 sheets, infructescence, leaves); Idem, NE Somalia, environs de Galgala,
C. Barbier, 962, 5.xii.1983 (K) (infructescence, basal bracts); Idem, Royal Botanical Gardens, Kew, (ex
horto), leg. P. Wilkin, A. Marrero & R.S. Almeida, 21.x.1997 (LPA: 18509) (leaves). D. hanningtoni Baker,
E Trop. Africa, German East Africa, Unyamwezi, Msalala, coll. & com. rev. J. Hannington (K TYPUS)
(inflorescence).
