ArticlePDF Available

Phylogeny and systematics of Lithodora (Boraginaceae—Lithospermeae) and its affinities to the monotypic genera Mairetis, Halacsya and Paramoltkia based on ITS1 and trnLUAA-sequence data and morphology


Abstract and Figures

Lithodora (Boraginaceae tribe Lithospermeae) comprises nine species of often narrowly endemic dwarf shrubs and shrubs. The genus shows a Mediterranean distribution with the centre of diversity in the western Mediterranean region. Lithodora is distinguished from its allies in Lithospermeae by a shrubby habit, the absence of both faucal scales and an annulus, and a very unusual mericarpid morphology, and it has therefore been regarded as a natural group in the past. Phylogenetic relationships of Lithodora were investigated with nuclear ribosomal ITS1 and chloroplast trnLUAA intron sequences using parsimony and likelihood analyses. The results strongly indicate a polyphyly of Lithodora, which falls into two clades: The majority of species forms a monophyletic group ("Lithodora II") in a well supported clade together with a Lithospermum s.l. clade (including the New World genera Onosmodium and Macromeria) and our sample of Buglossoides species. A smaller number of species ("Lithodora I") falls into a second well supported clade including the monotypic genera Mairetis, Paramoltkia and Halacsya. A critical reevaluation of carpological characters traditionally used to delimit Lithodora corroborates the results of the molecular analysis and indicates that the cupulate areole and the mericarpid appendage found in the genus are homoplasious. The polyphyly of Lithodora is formally recognised by removing "Lithodora II" as a newly established genus, Glandora. The seven newly required combinations in the genus are provided.
Content may be subject to copyright.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
Boraginaceae tribe Lithospermeae (DC.) Gürke are
an apparently monophyletic group (Hilger & Böhle, 2000;
Långström & Chase, 2002) with centres of diversity in
the Mediterranean basin, and the United States and
Mexico. Fruit anatomy indicates that tribe Lithosper-
meae is closely allied to Echiochilon Desf., Ogastemma
Brummitt, Antiphytum DC. ex Meisn. and Sericostema
Stocks ex Wight (Seibert, 1978) and it is now understood
to represent the sister group to tribe Echiochileae (Riedl)
Långström & M.W. Chase (Långström & Chase, 2002).
The systematics of tribe Lithospermeae was last studied
in detail by Johnston (1924, 1952, 1953a, b, 1954a, b),
who recognised several small segregate genera of Litho-
spermum L. In the Mediterranean region these include
the monotypic genera Mairetis I.M. Johnst., Moltkiop-
sis I.M. Johnst., and Neatostema I.M. Johnst., and the
slightly larger genera Buglossoides Moench (ca. ten spe-
cies), and Lithodora Griseb. (nine species). Delimitation
of Lithospermeae has been anything but straightforward
and there has been considerable confusion with regards to
the systematics of some taxa placed into Lithospermeae
by Johnston (1924, 1953a). Johnston (1954b) himself re-
moved some of the genera he had originally described
as segregates from Lithospermum (Mairetis, Moltkiop-
sis, Neatostema) to Eritricheae (Benth. & Hook.) Gürke.
Riedl (1967, 1968) subsequently moved these genera to
a new subtribe Moltkiopsidinae Riedl in his tribe Trigo-
notideae ( Popov) Riedl. However, Riedl (1967, 1968) also
included taxa in Trigonotideae which have since been
shown to belong to Cynoglosseae DC. s.l. (Tr igon otis
Stev.) and Echiochileae (Antiphytum, Sericostoma) re-
spectively (ngström & Chase, 2002). Seibert (1978)
already expressed serious doubts about the justification
of “Trigonotideae” in general and argued for an affinity
of Mairetis, Moltkiopsis and Neatostema to Lithosper-
meae on the basis of a detailed study of fruit morphology
and anatomy. This view was, however, not adopted by
Riedl (1997), who retained them in “Trigonotideae”, but
reluctantly admitted palynological similarity to Litho-
spermeae. Lithodora was universally placed in Litho-
spermeae (Joh nston, 1953b, 1954a; Seibert, 1978; Bigazzi
& Selvi, 1998), but Johnston (1924) and Melchior (1964)
emphasized a doubtful affinity to Boragineae Bercht. &
J. Presl. Riedl (1968) also emphasized that because of its
Phylogeny and systematics of Lithodora (Boraginaceae—Lithospermeae)
and its affinities to the monotypic genera Mairetis, Halacsya and
Paramoltkia based on ITS1 and trnLUAA-sequence data and morphology
Daniel C. Thomas1,2, Maximilian Weigend2 & Hartmut H. Hilger2
1 Royal Botanic Garden Edinburgh, 20a Inverleith Row, Edinburgh, EH3 5LR, UK.
(author for correspondence)
2 Freie Universität Berlin, Institut für Biologie – Systematische Botanik und Pflanzengeographie,
Altensteinstr. 6, 14195 Berlin, Germany
Lithodora (Boraginaceae tribe Lithospermeae) comprises nine species of often narrowly endemic dwarf
shrubs and shrubs. The genus shows a Mediterranean distribution with the centre of diversity in the western
Mediterranean region. Lithodora is distinguished from its allies in Lithospermeae by a shrubby habit, the
absence of both faucal scales and an annulus, and a very unusual mericarpid morphology, and it has therefore
been regarded as a natural group in the past. Phylogenetic relationships of Lithodora were investigated with
nuclear ribosomal ITS1 and chloroplast trnLUAA intron sequences using parsimony and likelihood analyses.
The results strongly indicate a polyphyly of Lithodora, which falls into two clades: The majority of species
forms a monophyletic group (“Lithodora II”) in a well supported clade together with a Lithospermum s.l. clade
(including the New World genera Onosmodium and Macromeria) and our sample of Buglossoides species. A
smaller number of species (“Lithodora I”) falls into a second well supported clade including the monotypic
genera Mairetis, Paramoltkia and Halacsya. A critical reevaluation of carpological characters traditionally used
to delimit Lithodora corroborates the results of the molecular analysis and indicates that the cupulate areole
and the mericarpid appendage found in the genus are homoplasious. The polyphyly of Lithodora is formally
recognised by removing “Lithodora II” as a newly established genus, Glandora. The seven newly required
combinations in the genus are provided.
KEYWORDS: Glandora gen. nov., ITS1, Lithodora, Mediterranean, phylogeny, trnLUAA intron
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
“shrubby habit” Lithodora represented “a primitive type”
in Lithospermeae.
Lithodora comprises nine species (plus three infra-
specific taxa variously considered as distinct species), all
of which are dwarf shrubs or shrubs (Fernandes, 1972).
Lithodora has a Mediterranean distribution with the cen-
tre of diversity in the western Mediterranean (Meusel &
Kästner, 1990). Several taxa are narrowly endemic (see
Fig. 1): L. moroccana I.M. Johnst. is endemic to north-
eastern Morocco, L. nitida (Ern) R. Fern. to a few moun-
tain ranges in Andalusia (southern Spain), L. oleifolia
(Lapeyr.) Griseb. to a small area in the eastern Pyrenees
(northeastern Spain), and L. zahnii (Heldr. ex Halácsy)
I.M. Johnst. to a small area of the western Peloponnese
The genus Lithodora was created by Grisebach (1844)
to accommodate six shrubby Mediterranean species for-
merly assigned to Lithospermum. Grisebach (1844) dif-
ferentiated Lithodora from Lithospermum mainly on the
basis of its shrubby (versus herbaceous) growth form and
apparent differences in floral morphology. However, most
subsequent authors ignored his genus and the name was
considered a synonym of Lithospermum or was recognised
only at infrageneric level (e.g., Candolle, 1846; Boissier,
1879; Gürke, 1897), until Johnston published a detailed
analysis of Lithodora (Johnston, 1953b). He reinstated
Lithodora on the basis of its peculiar fruits, which he con-
sidered as “unique in the whole Boraginaceae” (Johnston,
1953b: 260). The mericarpids show a circumscissile abscis-
sion of the upper, seed-bearing part, while the lower part
remains connected with the gynobase forming a cupulate
structure (Johnston, 1953b). In addition to the unique fruit
morphology, Johnston (1953b) found several other char-
acters distinguishing Lithodora from Lithospermum s.str.
Fig. 1. Distribution map of Halacsya, Lithodora, Mairetis and Paramoltkia. Based on the distribution map in Meusel & al.,
1978, vouchers examined in B, BSB, E, FR, and MSB and the following publications: Meusel & Kästner, 1990 (Lithodora);
Stevanović & al., 2003 (Halacsya, Paramoltkia); Johnston, 1953b; Sauvage & Vindt, 1954; Silva & Rozeira, 1962; Valdés,
1981; Browicz, 1985; Loidi & al., 1988; Bolòs i Capdevila & al., 2001; Tan & Iatrou, 2001; Valdés & al., 2002; and Bañares
Baudet, 2003.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
such as the shrubby growth form and the absence of both
corolla invaginations (faucal scales) and basal scales (an
annulus) in the corolla tube. Johnston (1953b) also created
three sections in Lithodora, mainly based on the occur-
rence of different stylar polymorphisms (absent, stigma-
height dimorphism, distyly) and the morphology of the
detachment scar of the mericarpid ( = cicatrix) and the
detachment scar at the receptaculum ( = areole). Section
Lithodora (only L. fruticosa (L.) G riseb.) is ch aracte rised
by a stigma-height dimorphism, i.e., only the length of the
styles vary between the morphs, but the level of filament
attachment and the filament length are relatively con-
stant between the morphs. The mericarpids have a broad
appendage, which does not include a sclerenchymatous,
tubular channel. Section Allostema I.M. Johnst. (L. olei-
folia, L. moroccana, L. rosmarinifolia (Ten.) I.M. Johnst.,
L. hispidula (Sibth. & Sm.) Griseb., L. zahnii ) is charac-
terised by distylous flowers and a mericarpid append-
age which includes a sclerenchymatous tubular channel.
According to Johnston (1953b), section Lasioglottis I.M.
Johnst. is characterised by monomorphic flowers and fila-
ments which are affixed at unequal heights on the corolla
tube (only L. prostrata (Loisel.) Griseb.), but Fernandes
(1970) and Valdés (1981) observed that L. prostrata ex-
hibits true distyly and can thus not be differentiated from
section Allostemma. Seibert (1978) doubted Johnston’s
theory of a suprabasal abscission of the mericarpids in
his very informative carpological overview of the Litho-
spermeae, thereby weakening this apparent synapomor-
phy of the Lithodora species to a certain extent. However,
he still agreed with Johnston on the basic uniformity of
the mericarpids within Lithodora, and he quoted Sprengler
(1919: 115), who had emphasized that Lithodora species
exhibit “not only great similarities in floral morphology
(minor differences of form and size of the corolla), but also
in vegetative morphology” (own translation).
The flowers of Lithodora show conspicuous charac-
ters that differentiate its members from Lithospermum
s.str. and/or Buglossoides: the corollas in Lithodora are
blue or purple in contrast to the yellow, orange or whitish
corollas predominant in Lithospermum s.str. They lack
both faucal scales and an annulus (Sprengler, 1919; John-
ston, 1953b), while corolla invaginations are present in
most species of Lithospermum s.str. (faucal scales) and
Buglossoides (vertical pleats), and almost all species in
these two genera exhibit an annulus (Johnston, 1954a;
Ralston, 1993). However, other genera of Lithospermeae
also exhibit a shrubby or subshrubby habit and blue corol-
las (e.g., Moltkia Lehm.), and an annulus and faucal scales
are missing in a range of taxa in the Lithospermeae, for
example in Moltkia as well as in Paramoltkia Greuter. In
Arnebia Forssk. faucal scales are always missing while
an annulus is developed in some species, but missing in
others. Overall, none of the morphological characters de-
fining and delimiting the genus Lithodora sensu John-
ston (1953b) is unique in Lithospermeae, with the only
exception of the peculiar fruit morphology. On the basis
of the apparent overall similarity of the taxa reunited in
Lithodora and also on the basis of its apparently unique
carpological condition as identified by Johnston (1953b)
the genus Lithodora became a widely recognised segre-
gate of Lithospermum (Silva & Rozeira, 1962; Fernandes,
1971, 1972; Valdés, 1981; Meikle, 1983; Greuter & al.,
1984; Browicz, 1985; Al-Shehbaz, 1991; Riedl, 1997).
The present study initially intended to clarify the
systematic position of Lithodora sensu Johnston (1924,
1953b) in the tribe Lithospermeae. However, preliminary
molecular data surprisingly associated different species
of the genus with different clades in Lithospermeae, indi-
cating a relationship of some Lithodora species with the
monotypic Balkan endemics Halacsya Dörfl. and Para-
moltkia, which are either considered to be closely allied
to Moltkia (Johnston, 1953b; Stevanović & al., 2003) or
to Echium L. (Seibert, 1978; Greuter, 1981). As has been
indicated above, there are thus several open questions
with regards to the systematics of Lithospermeae, which
require clarification: Is Lithodora in the circumscription
of Johnston (1953b) monophyletic? What are its affini-
ties? Do vegetative and generative characters which were
traditionally used to delimit Lithodora, especially the
carpological character set, represent autapomorphies of
a natural group? Can the shrubby growth habit in Litho-
dora be interpreted as indicative of “a primitive type” in
Lithospermeae? What are the affinities of the obligate
serpentine endemics Halacsya and Paramoltkia? Do the
monotypic genera Mairetis, Moltkiopsis and Neatostema
belong to Lithospermeae? The findings will then be re-
flected in a —if necessarymodifie d classif ication of the
genus Lithodora.
Sampling. —
Thirty-eight ingroup taxa were incor-
porated in the molecular matrix including the nine spe-
cies of Lithodora sensu Johnston (1953b), four species of
Buglossoides, and the segregate genera Neatostema and
Mairetis. Unfortunately, we did not succeed in obtaining
useful sequences of Moltkiopsis. The large genus Litho-
spermum s.str. is represented by eight taxa representing
all growth habits known in the genus (herb, subshrub,
dwarf shrub and shrub) and by representatives from both
the Old World and the New World in order to provide a
preliminary test of its monophyly. The majority of genera
associated with Lithospermeae is included in the analy-
sis. We focussed on Old World genera with the exception
of Lithospermum s.str. (sensu Johnston, 1954a), which
is the only genus of the Lithospermeae present in both
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
the Old and the New World, and the closely allied North
American genera Macromeria D. Don and Onosmodium
Michx. Representatives of the tribes Echiochileae and
Boragineae were chosen as outgroup taxa on the basis of
the study by Långström & Chase (2002) and the observa-
tions of Seibert (1978). Voucher information, collection
and herbarium details, and GenBank accession numbers
of the taxa used for the molecular analysis are presented
in the Appendix.
DNA extraction, amplification and sequenc-
ing. —
DNA was extracted from fresh or silica-gel dried
leaves or leaves of herbarium material, as described by
Weigend & al. (2004). The primers P1 and P2 (Baldwin,
1992) were used for the PCR-fragment amplification of
the ITS1 region, and the primers C and D (Taberlet & al.,
1991) for the amplification of the trnLUAA intron. Ampli-
fied products were purified using QIAquick
according to the manufacturer’s protocol (Qiagen Inc.).
The purified amplification products served as templates
for cycle sequencing with a Quick Start Kit (Beckman
Coulter) according to manufacturers instructions. Se-
quencing of the ITS1 region and the trnLUAA intron was
done using the same primers as in the amplification. A
CEQ 8000 (Beckman Coulter) automated sequencer was
used to generate the data. Sequences of Echium vulgare
L. and E. wildpretii H. Pearson were taken from the study
by Böhle & al. (1996), sequences of Anchusa officinalis
L. and Borago officinalis L. were taken from the study
by Hilger & al. (2004).
Editing, alignment and gap coding. —
The initial
sequence data were edited using Chromas Pro V. 1.33
(Technelysium Pty. Ltd. 2003–2005). The alignment was
performed manually in Multicolor Sequence Alignment
Editor (Hepperle, 2001). A secondary structure model of
the ITS1 transcript has been proposed for Boraginales
(Gottschling & al., 2001). This model consists of an open
multibranch loop with four pairing regions (helices I to
IV). The alignment of the ITS1 sequences proved diffi-
cult in the part corresponding to helix I of the transcript.
Therefore, a sequence of 39 chara cters was excluded from
the ITS1 matrix. The rest of the alignment was aided by
the investigation of the framework dictated by the second-
ary structure of the ITS1 transcript, with separate align-
ing of homologous regions that were identified by mut ual
sequence comparison. Gaps were coded as missing data,
but eleven indels in the ITS1 region and three indels in
the trnLUAA intron were coded as additional parsimony-
informative binary characters (present/absent). The align-
ment is available on TreeBASE.
Phylogenetic analysis. —
Phylogenetic relationships
were analysed using winPAUP 4.0b10 (Swofford, 2002).
To assess character congruence between the ITS1 and the
datasets, a partition homogeneity test (Farris & al.,
1994) as implemented in winPAUP 4.0b10 was performed
with 1,000 replicates with no more than 1,000 trees saved
per replicate. Heuristic parsimony searches of combined
nrDNA and cpDNA datasets were performed with 1,000
random sequence additions
all characters unweighted and
unordered, and the following options in effect: tree bisec-
tion-reconnection (TBR) branch swapping, collapse zero
length branches, and MULTREES. Confidence limits for
tre es were assessed by perfor ming 1,000 repl icates of boot-
strapping (Felsenstein, 1985) with 100 addition sequence
replicates per bootstrap replicate using equal weighting,
TBR branch swapping and MULTREES on. Further sup-
port was estimated using jackknife analysis. We ran 1,000
replicates with 37% deletion, using random addition se-
quence (10 replicates) and TBR branch swapping.
Maximum likelihood calculations were executed as-
suming a GTR + G + I model and a rate variation among
sites following a gamma distribution. GTR + G + I was
chosen as the best-fit substitution model using the hi-
erarchical likelihood ratio test (hLRT) in Modeltest 3.7
(Posada & Crandall, 1998). The settings proposed by
Modeltest were: Basefreq = (0.3589 0.1687 0.1798), Nst =
6, Rmat = (1.0000 1.3194 1.0000 1.0000 2.7946), Shape
= 0.6277, Pinvar = 0.3470. Likelihood bootstrap searches
were performed with 1,000 replicates using the “fast boot-
strap” option (bootstrapping without branch-swapping).
“Fast bootstrapping” was shown to provide estimates of
support similar to, though generally less than, bootstrap-
ping with branch-swapping within a reasonable comput-
ing time (Mort & al., 2000).
Scanning electron microscopy. —
material for scanning electron microscopy (SEM) was
di rectly mou nted on double-sided adhesive tap e and sput-
ter coated with gold. Material from the FAA collection
was first dehydrated in a graded ethanol series, critical
point dried following the standard procedure, and than
mounted and sputter coated as above. The material was
subsequently viewed and photographed with a LEO VP
430 scanning electron microscope.
Molecular phylogeny. —
The partition homogene-
ity test indicated that our data from two distinct marker
regions were not significantly different from random par-
titions of the combined dataset (P = 0.46; Cunningham,
1997), thereby justifying the combined analysis of both
loci. The combined matrix of ITS1 and trnL
(group I) data contained 42 taxa and 722 characters (in-
cluding 14 indel codes). Of these 252 (11 indels) were
derived from the ITS1 dataset and 470 (3 indel codes)
from the trnLUAA intron dataset.
Analyses of the combined dataset yield trees with
similar overall topologies in both parsimony and maxi-
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
mum likelihood searches. The heuristic search produced
5,520 shortest trees of 536 steps with a consistency index
(CI) of 0.63 and a retention index (RI) of 0.73. Figure 2
shows the strict consensus tree of the 5,520 most parsi-
monious trees with bootstrap support values above the
branches and jackknife values below. Figure 3 shows
the single optimal tree obtained by maximum likelihood
analysis of t he combined dat a with bootstrap suppor t val-
ues above the branches or to the right of nodes. Bootstrap
(bs) and jackknife support values ( jk) given in brackets
in the following text are derived from the parsimony
analysis. The ingroup, i.e., all analysed Lithospermeae
taxa, is strongly supported as monophyletic (bs: 97, jk:
Fig. 2. Parsimony analysis of combined ITS1 and trnLUAA data. Strict consensus tree of the 5,520 most parsimonious trees
(length = 536, CI = 0.63, RI = 0.73) of the heuristic search. Bootstrap support values > 50% are presented above the branch-
es, jackknife values > 50% below. BOR, Boragineae; ECH, Echiochileae; LIRA I, “Lithodora I”; LIRA II, “Lithodora II”.
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
Lithodora falls into two monophyletic assemblages:
Lithodora I” (bs: 84, jk: 90) comprises L. hispidula, L.
zahnii, and L. fruticosa. Within this clade L. fruticosa is
the sister taxon to a strongly supported L. zahnii/L. hispi-
dula clade (bs: 100, jk: 100). “Lithodora I” is a subclade of
a moderately supported clade A (bs: 87, jk: 91), which also
contains Paramoltkia doerfleri (Wettst.) Greuter & Bur-
det, Halacsya sendtneri (Boiss.) Dörfl. and Mairetis mi-
crosperma (Boiss.) I.M. Johnst. The relationships among
Lithodora I” and the other taxa in clade A are only poorly
resolved, but there is weak support for a sister group re-
lationship of this clade to Neatostema apulum (L.) I.M.
Johnst. (bs: 77, jk: 84). Moltkiopsis ciliata (which we failed
to sequence successf ully) will a lmost cer ta in ly be ret rieve d
in this clade. Its presence or absence does not modify the
morphological or phylogenetic interpretations, since it is
Fig. 3. Likelihood analysis of combined ITS1 and trnLUAA data. Single optimal likelihood tree obtained by a heuristic search
using the GTR + G + I substitution rate matrix, plus a gamma distribution of rate variation among sites and a proportion of
invariant sites (settings: Basefreq = [0.3589 0.1687 0.1798], Nst = 6, Rmat = [1.0000 1.3194 1.0000 1.0000 2.7946], Shape =
0.6277, Pinvar = 0.3470). Bootstrap support values generated by “fast bootstrapping” are shown above the branches or to
the right of nodes (percentages < 50% not shown). Broken lines indicate branches with < 50% bootstrap support.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
morphologically similar and evidently closely related to
Neatostema and Mairetis (Joh nston, 1953a, 1954b; Riedl,
1967, 1968). “Lithodora II” forms a strongly supported
clade (bs: 95, jk: 97), comprising L. nitida, L. oleifolia,
L. moroccana, L. rosmarinifolia, L. prostrata and L. dif-
fusa (Lag.) I.M. Johnst. Within this clade L. nitida and L.
oleifolia form a weakly supported sister group to the other
species in this clade (not resolved in the strict consensus
tree of the MP analysis). The relationships among these
species are well resolved: Lithodora moroccana forms a
well supported clade with L. rosmarinifolia (bs: 91, jk:
94), and L. prostrata forms a moderately supported clade
with L. diffusa (bs: 88, jk: 87). “Lithodora II” is retrieved
as monophyletic group within a well supported clade B
(bs: 93, jk: 96) together with a Lithospermum s.l. clade
(bs: 91, jk: 96, including Macromeria and Onosmodium)
and our sample of Buglossoides species. Buglossoides sec-
tion Buglossoides is strongly supported as monophyletic
(bs: 100, jk: 100), however, the monophyly of the genus
Buglossoides is not supported. A sister group relationship
between Buglossoides /Lithospermum on the one hand and
Lithodora II” on the other is only weakly supported (bs:
74, jk: 79).
Morphology. —
Selected vegetative and generative
characters of twelve species and subspecies of Lithodora
and of Neatostema, Mairetis, Moltkiopsis, Halacsya and
Paramoltkia are summarized in Table 1. Selected veg-
etative characters and floral characters are illustrated
in Fig. 4. Figures 5 and 6 illustrate systematically im-
portant morphological characters of the mericarpids
Table 1. Selected morphological characters of Lithodora and allied taxa (continued on next page).
Clade Taxon
Growth form
Special under-
ground organs
Leaf (lamina
Leaf (indu-
Style polymor-
A, L I Lithodora fruticosa DS E L, Obl, E S-H C S S
A, L I Lithodora hispidula subsp.
hispidula DS E Obov S-H S, (C) S D
L I Lithodora hispidula subsp.
cyrenaica DS, S E Obov S-H dT, C S D
L I Lithodora hispidula subsp.
versicolor DS, S E Obov S-H dT, C S D
A, L I Lithodora zahnii DS E L, Obl S dT, C S D
AParamoltkia doerfleri HS E R Obl, E S C SP
AHalacsya sendtneri HS A, E Obov, Obl, Ov S C SP
AMairetis microsperma A A Obov, Obl S C S, SP
ANeatostema apulum A E Obov, L, Ov S C, T S, SP
Moltkiopsis ciliata HS A, E Ov, Obl S-H C S
B, L II Lithodora nitida DS A, D S Obov, E S C S D
B, L II Lithodora oleifolia DS A, D S Obov, E S-T C P D
B, L II Lithodora moroccana DS A, D Obov, Obl S-T C S D
B, L II Lithodora rosmarinifolia DS A, E L, Obl, E S C S D
B, L II Lithodora prostrata subsp.
prostrata DS A, D L, Obl, E H, S C S D
L II Lithodora prostrata subsp.
lusitanica DS A, E L, O, E H, S C S D
B, L II Lithodora diffusa DS A, D L, O, E H, S C S D
Clade: A, clade A; B, clade B (L I, “Lithodora I”; L II, “Lithodora II”; see Fig. 2). Growth form: A, annual; DS, dwarf shrub;
HS, subshrub; S, shrub. Habit: A, ascending; D, decumbent; E, erect. Special underground organs: –, no special under-
ground organs present; R, rhizome; S, perennial underground stolons. Leaf (lamina form): E, narrowly elliptic; L, linear; Obl,
narrowly oblong; Obov, obovate or narrowly obovate; Ov, narrowly ovate. Leaf (indumentum): H, hispid; S, strigose; S-H,
strigose-hispid; S-T, strigose-tomentose. Inflorescence: C, cymoid; dT, depauperate thyrsoids; S, flowers solitary (depauperate
cymoids); T, thyrsoids. Pedicellus: S, flowers sessile or subsessile; P, distinctly pedicellate; SP, shortly pedicellate. Style
polymorphisms: –, absent; D, distyly; S, stigma-height dimorphism.
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
and the receptaculum: the cicatrix ( = detachment scar
at the mericarpid), the cicatrix appendage, the tubular
funicular channel, and the areole ( = detachment scar at
the receptaculum).
Fruit morphology. —
The terminology of this sec-
tion follows Seibert (1978) and Hilger (1985). The meri-
carpids in Lithodora are ovoid, and straight, or slightly
(L. oleifolia) to strongly (L. hispidula, L. zahnii) incurved
towards the style. They are gradually narrowed to the
base or suprabasally constricted and then exhibiting a
thicker collar about the base (L. prostrata, L. fruticosa)
or a short, stout neck (L. hispidula, L. zahnii). The ven-
tral keel is protruding, at least in the distal portion. The
mericarpids are dorsally convex, a dorsal keel is missing
or weakly developed in the distal part (L. fruticosa, L.
zahnii, L. prostrata subspp.). The mericarpids are whit-
ish, grey, brownish, or black. The surface is smooth, or
slightly tumulose (Lithodora nitida), minutely tuberculate
(L. prostrata subspp.), muriculate (L. hispidula subspp.),
or minutely rugulose (L. fruticosa, L. zahnii). The cica-
trix is horizontal to slightly oblique. All species except
L. nitida show a distinct peg-like or broadly pyramidal
to asymmetrical cicatrix appendage in central to slightly
ventral position. Around this appendage, the cicatrix is
often distinctly hollowed-out by the collapse of the ab-
scission tissue (L. diffusa, L. moroccana, L. oleifolia, L.
prostrata subspp.) or the cicatrix around the appendage
is only weakly sunken (L. rosmarinifolia, L. hispidula,
Table 1. Continued.
Clade Taxon
Petal colour
hairs (throat)
(corolla tube)
Stamina inser-
A, L I Lithodora fruticosa P, B–––EFCTR–
A, L I Lithodora hispidula subsp.
hispidula B, P E F CT M
L I Lithodora hispidula subsp.
cyrenaica B–––EFCTM?
L I Lithodora hispidula subsp.
versicolor B, P, R E F CT M ?
A, L I Lithodora zahnii B, (W) E F CT R
AParamoltkia doerfleri P–––E–FS–
AHalacsya sendtneri Y–––E–FT, T-R
AMairetis microsperma B + Y + E F T +
ANeatostema apulum Y++E FT+
Moltkiopsis ciliata B, P + + D F S, T +
B, L II Lithodora nitida B + + E VD STu ?
B, L II Lithodora oleifolia B–++EPDS+
B, L II Lithodora moroccana B, (W) + + E P D S ?
B, L II Lithodora rosmarinifolia P, B, (W) + + E P D S +
B, L II Lithodora prostrata subsp.
prostrata B, (P) + + D P D MT
L II Lithodora prostrata subsp.
lusitanica B, (P) + + D P D MT ?
B, L II Lithodora diffusa B–++EPDS?
Clade: A, clade A; B, clade B (L I, “Lithodora I”; L II, “Lithodora II”; see Fig. 2). Petal colour: B, blue; P, purplish; R, pink-
ish; W, white; Y, yellow. Annulus (basal scales): – , absent; +, present. Glandular hairs (adaxial corolla tube, corolla throat):
, absent; +, present. Indument (abaxial corolla tube): – , absent, +, present. Stamen insertion: D, insertion on different lev-
els (for details, see Johnston, 1953a, b); E, stamen insertion on one level. Cicatrix appendage: – , absent or minutely protruding
vascular strands or funicular channels; F, Lithodora fruticosa-type; P, Lithodora prostrata-type. Areole: CT, cupulate, thin-
walled; D, cupulate, thick-walled depression in the receptacle; F, flat, slightly concave or with minutely elevated margin; VD,
oblique with ventral depression. Mericarpid surface: M, muriculate; MT, minutely tuberculate; R, rugulose; S, smooth; STu,
slightly tumulose, T, tuberculate; T-R, tuberculate-rugose. CaCO3 incrusts in pericarp: data from Seibert (1978), – , absent; +,
present; ?, unknown. Abbreviations in brackets indicate rare character states.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
Fig. 4. Growth habit and floral morphology in Lithodora. A, L. hispidula subsp. hispidula (A. Kagiampaki 9–00), habit; B,
L. zahnii (D. Thomas 2–05), habit. C–E, L. oleifolia (D. Thomas 3–05). C, habit; D, habit, note perennating underground
runners; scale bar = 10 cm; E, flower. F, L. prostrata (D. Thomas 10–05), flowers; G, L. hispidula subsp. hispidula (A. Ka-
giampaki 9–00), flower; H, L. fruticosa (D. Thomas 7–06), flower. I, L. zahnii (D. Thomas 2–05), flowers.
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
L. zahnii ) or flat (L. nitida, L. fruticosa). The areoles are
distinctly cupulate in all species except L. nitida, which
shows only a weakly developed depression in ventral posi-
tion of the oblique areole.
In spite of ma ny shared characters in the morphology
of the mericarpids, two distinct types can be differenti-
ated on the basis of the morphology of the cicatrix and
the areole:
1. Lithodora fruticosa-type: The mericarpids show
a centrally to slightly ventrally positioned, broadly py-
ramidal to asymmetrical appendage (Fig. 5A–C). In con-
trast to the cicatrix appendages of the L. prostrata-type
the appendage of the L. fruticosa-type does not include
a sclerenchymatous, tubular channel. However, a scle-
renchymatous tubular channel leaves the mericarpid in
ventral position directly beneath the keel. The areoles are
Fig. 5. Scanning electron micrographs of mericarpids and areoles: Lithodora fruticosa-type. A, L. fruticosa (D. Thomas
7–06), mericarpid; B, L. hispidula subsp. hispidula (A. Kagiampaki 9–00), mericarpid; C, L. zahnii (D. Thomas 2–05), meri-
carpid. D–E, L. fruticosa (D. Thomas 7–06). D, mericarpid, attached; E, cupulate areole. F–G, L. hispidula subsp. hispidula
(A. Kagiampaki 9–00). F, mericarpid, attached; G, cupulate areole. Abbreviations: am, abortive mericarpid; ca, cupulate
areole; cx, cicatrix appendage; mc, mericarpid; s, style; vs, vascular strand. Scale bars = 1 mm.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
characterised by a hard, thin-walled, cupulate structure. In
contrast to t he depression of t he receptacu lum cha racteris-
tic of the L. prostrata-type, the walls of the cupulate areole
reach a height of up to ca. 1 mm above the attachment level
of the abortive mericarpids (Fig. 5D–G). A sclerenchy-
matous tubular channel filled by a vascular strand runs
through the ventral wall of the areole and is recognisable
as a ridge on the inner wall. It slopes down the wall and
inwards at the bottom of the cupulate areole (Fig. 5E, G).
This channel is a continuation of the channel which leaves
the mericarpid in ventral position directly below the keel
(L. fruticosa, L. hispidula subspp., L. zahnii ).
Fig. 6. Scanning electron micrographs of mericarpids and areoles: Lithodora prostrata-type. A, L. prostrata (T. Raus
20775), mericarpid; B, L. diffusa (D. Thomas 5–05), mericarpid; C, L. rosmarinifolia (E. di Gristina 15.02.03), mericarpid; D,
L. oleifolia (H. Hilger 25.5.04), mericarpid; E, L. nitida (D. Thomas 1–05), mericarpid; F, L. diffusa (D. Thomas 5–05), areole;
G, L. prostrata (T. Raus 20775 ), mericarpid, still attached; areole; H, L. diffusa (D. Thomas 5–05), mericarpid, still attached;
areole; I, L. nitida (F. Garcia & al. 2378/79), areole. Abbreviations: am, abortive mericarpid; ca, cupulate areole; cx, cicatrix
appendage; mc, mericarpid; oa, oblique areole; s, style; vs, vascular strand. Scale bar = 1 mm.
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
2. Lithodora prostrata-type: The mericarpids show
a ventrally to centrally positioned, peg-like cicatrix ap-
pendage mainly consisting of a sclerenchymatous, tubu-
lar channel (Fig. 6A–D). After detachment this channel
is recognisable as a hole at the proximal end of the ap-
pendage. In L. nitida the “appendage” is only very weakly
developed and the sclerenchymatous funicular channel
is only slightly protruding in ventral position (Fig. 6E).
The areoles are cup-shaped, thick-walled depressions in
the receptaculum in all species in this group except of
L. nitida (Fig. 6F–H). In that latter species the oblique
areole exhibits only a weakly developed depression in
ventral position (Fig. 6I). The remains of the vascular
strand, which occupied the tubular channel in the cica-
trix appendage before the mericarpid detachment, arise
bristle-like in ventral to central position from the bottom
of the cupulate areole (L. diffusa, L. moroccana, L. nitida,
L. oleifolia, L. prostrata subspp, L. rosmarinifolia) (Fig.
Johnston’s interpretation of the carpological condi-
tions in Lithodora as unique in the whole Boraginaceae
and as crucial differential characters of the genus (John-
ston, 1953b) is recognised as erroneous on the basis of
the data presented here. Lithodora is polyphyletic and
falls into two clades: “Lithodora I”, which comprises
the western Mediterranean L. fruticosa, and the eastern
Mediterranean species L. hispidula and L. zahnii. This
clade is closely allied with the three monotypic, Medi-
terranean genera Mairetis, Paramoltkia and Halacsya.
Lithodora II”, which comprises the western Mediter-
ranean species L. nitida, L. oleifolia, L. moroccana, L.
rosmarinifolia, L. prostrata and L. diffusa, and is closely
allied to Lithospermum s.str. and Buglossoides. The two
clades here retrieved do not coincide with the infrageneric
classifications of Lithodora by Johnston (1953b). Section
Allostema is polyphyletic as L. hispidula and L. zahnii are
retrieved in “Lithodora I”, while L. oleifolia, L. moroc-
cana and L. rosmarinifolia are retrieved in “Lithodora
II”. The molecular data receive clear support from the
reexamination of carpological characters: “Lithodora I” is
characterised by a broad cicatrix appendage, which does
not include a sclerenchymatous tubular channel. However,
a sclerenchymatous tubular channel leaves the mericarpid
in ventral position directly beneath the keel and continues
into the ventral wall of thin-walled cupulate areole (L.
fruticosa-type, Fig. 5). In contrast to the depressions of the
receptacle found in “Lithodora II”, the cupulate areoles in
Lithodora I” are formed by a hard, thin-walled structure,
which rises distinctly above the attachment level of the
abortive mericarpids. Johnston (1953b) already described
this syndrome in his treatise of Lithodora as an aberrant
condition found only in L. fruticosa. However, neither he
nor Seibert (1978) recognised the similarity of this syn-
drome to the condition found in L. hispidula and L. zahnii
and the taxonomic value of this character set. “Lithodora
II” is characterised by a peg-like cicatrix appendage which
mainly consists of a sclerenchymatous, tubular channel.
The areoles are thick-walled depressions of the receptacle
and show the remains of the vascular strand, which occu-
pied the tubular channel in the cicatrix appendage before
the detachment of the mericarpid (L. prostrata-type, Fig.
6). The morphology of the mericarpids and areoles of L.
nitida is exceptional. This species exhibits only a weakly
developed, ventrally positioned mericarpid “appendage”
and an oblique cicatrix with a weakly developed depres-
sion in ventral position. This mericarpid morphology re-
sembles the conditions found in Lithospermum s.str. and
might represent the plesiomorphic condition.
Floral morphology provides some additional charac-
ters differentiating the two clades of Lithodora. All species
of Lithospermum s.str., many species of Buglossoides and
all species in “Lithodora II” share the presence of three-
celled gland-tipped trichomes in the corolla throat. These
three-celled glandular trichomes are absent in “Lithodora
I” and in the closely allied species Paramoltkia doerfleri,
Mairetis microsperma and Halacsya sendtneri. Moreover,
the corolla tubes of all species in “Lithodora II” are abaxi-
ally pubescent, as are the corollas in Lithospermum s.str.
and Buglossoides, whereas the corolla tubes of species of
Lithodora I”, and Paramoltkia, Halacsya and Mairetis
are abaxially glabrous.
Lithodora falls into two fairly derived clades in mostly
herbaceous Lithospermeae and its shrubby, evergreen
growth form clearly has to be considered as a derived
trait and likely a secondary adaptation to the dry Medi-
terranean climate. Woody chamaephytes have evolved
many times independently from herbaceous ancestors in
Boraginaceae, including several taxa within otherwise
herbaceous Lithospermum s.str. (e.g., Lithospermum
gayanum (Wedd.) I.M. Johnst.). The interpretation of a
shrubby habit as indicative of a “primitive type” (Riedl,
1968) clearly has to be rejected on the basis of phyloge-
netic data (see also Böhle & al., 1996 on Echium).
At first glance it seems sur prising that stylar polymor-
phisms are present in both “Lithodora I” and “Lithodora
II”, since they represent complex morphological and phys-
iological syndromes. However, styla r polymorph ism is al-
ready known to have evolved several times independently
within Lithospermeae (see review in Ganders, 1979a):
They are found in at least seven North American species
of Lithospermum s.str. (Johnston, 1952; Ganders, 1979b;
Ralston, 1993), and twelve species of Arnebia (John-
ston, 1952, 1954a, including Huynhia pulchra (Roem. &
Schult.) Greuter & Burdet). Together with the two inde-
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
pendent arisals in “Lithodora I” and “Lithodora II”, stylar
polymorphisms appear to have arisen at least four times
independently in this tribe alone (see Fig. 2). Stylar poly-
morphisms are also known in Boraginaceae tribe Borag-
ineae, e.g., in Anchusa L. (Philipp & Schou, 1981; Schou
& Philipp, 1984; Selvi & Bigazzi, 2003) and Pulmonaria
L. (Olesen, 1979; Richards & Mitchell, 1990), and also
in Boraginaceae tribe Eritricheae in several species of
Amsinckia Lehm. and in Cryptantha Lehm. ex G. Don
(Casper & al., 1988; Schoen & al., 1997; Li & Johnston,
2001). Boraginaceae thus represent an excellent model
group for the study of the evolution of stylar polymor-
phisms and correlated issues (breeding systems in sister
groups, life history traits) and our future research will be
directed towards addressing some of these questions.
Seibert (1978) and Hilger & al. (1993) observed that
calcium carbonate (CaCO3) incrustation in the outer
pericarp is characteristic of the Lithospermeae and of
related taxa, which are now placed in the Echiochileae
(Antiphytum, Echiochilon, Ogastemma, Sericostema).
This character might, therefore, represent a plesiomorphic
state in these two tribes. Seibert sampled 53 species in 23
genera in the Lithospermeae. Within his sample, calcium
carbonate incrustations were absent only in a few taxa:
Pontechium Böhle & Hilger, Echium L., Echiostachys
Levyns, and Lobostemon Lehm., all of which are closely
allied (“tribe Echieae Dumort.”, Hilger & Böhle, 2000;
Buys, 2006), and in “Lithodora I”, Halacsya and Para-
moltkia, which are retrieved in clade A in our molecular
analysis. Interestingly, calcium carbonate is also absent
in a single species of “Lithodora II”, L. prostrata, but
present in others (L. oleifolia, L. rosmarinifolia) and the
closely related genera Buglossoides and Lithospermum.
Calcium carbonate incrustation seems to have been lost
at least three times within Lithospermeae: in the Echium
alliance, in “Lithodora I” and closely allied taxa, and in
L. prostrata. In the latter species the lack of calcium car-
bonate in the pericarp might be correlated with edaphic
conditions. Lithodora prostrata prefers acidic soils, while
L. oleifolia and L. rosmarinifolia are calcicolous (Fern-
andes, 1972). Calcium carbonate incrustations are also
absent in both the calcicolous taxon “Lithodora I” and
the closely related obligate Balkan serpentine endemics
Halacsya and Paramoltkia. Close relationships between
calcicolous taxa and serpentine endemic taxa are typical
and have been reported for various Mediterranean plant
groups (Stevanović & al., 2003).
The phylogenetic data here presented surprisingly
confirm the systematic justification of most Old World
monotypic Lithospermeae genera recognised in the litera-
ture (Neatostema, Mairetis, Halacsya, Paramoltkia). The
segregation of Neatostema and Mairetis from Lithosper-
mum (Johnston, 1953b) and of Paramoltkia from Moltkia
(Greuter, 1981) on the basis of morphological characters
is vindicated by the molecular analysis presented here.
The genera Mairetis and Neatostema (and probably also
Moltkiopsis), which have been placed in Trigonotideae
(Riedl, 1967, 1968, 1997) and Eritricheae (Johnston,
1954b) in the past, are clearly nested in Lithospermeae
and form a well-supported, exclusively Mediterranean
clade together with Lithodora s.str., Halacsya and Para-
moltkia. These molecular data support the original clas-
sification of Johnston (1953a) and the fruit anatomical
data of Seibert (1978). Also, the removal of Paramoltkia
doerfleri from Moltkia is clearly supported. Paramoltkia
is not found to be closely associated to Moltkia at all, but
is allied with Halacsya. The fruit anatomical similarities
of Paramoltkia and Halacsya were already recognised
by Seibert (1978) and Greuter (1981). However, these au-
thors suggested a close affinity of these two genera to
Echium and Lobostemon, probably mainly based on the
lack of calcium carbonate incrusts in the pericarp in all
of these taxa and the zygomorphic flowers of Halacsya.
The results of our molecular analysis do not support this
placement, neither Halacsya nor Paramoltkia are closely
allied with Echium. The zygomorphic flowers of Halac-
sya arose independently of those of Echium from radially
symmetrical ancestors.
The molecular and the morphological data presented
here clearly necessitate taxonomic changes to accom-
modate the new phylogenetic insight of Lithodora sensu
Johnston (1953b) being polyphyletic. Since L. fruticosa is
the type of the generic name (Johnston, 1953b: 259) the
genus bearing that name has to be restricted to “Litho-
dora I” including L. fruticosa, L. hispidula and its sub-
species, and L. zahnii. This redefined genus Lithodora
is closely allied to the Mediterranean monotypic genera
Paramoltkia, Halacsya and Mairetis. Within this group
Lithodora I” is morphologically easily distinguishable by
its fruit morphology and growth form, while the whole
group is morphologically heterogeneous. Because of the
presence of a variety of autapomorphies in all of the mono-
typic genera (Johnston, 1953a, b; Fernandes, 1972; Greu-
ter, 1981) closely allied with “Lithodora I”, it would be
counter-productive to unite them and to create a morpho-
logically heterogeneous taxon solely based on molecular
data. “Lithodora II” is retrieved in a clade comprising the
large genus Lithospermum s.str., but also Buglossoides
and the North American segregates Macromeria and On-
osmodium. Extended sampling would very likely show
that all New World segregate genera (Lasiarrhenum I.M.
Johnst., Nomosa I.M. Johnst., Perittostema I.M. Johnst.,
Psilolaemus I.M. Johnst.) also fall i nto this latte r clade. To
accommodate the phylogenetic data in the classification
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
of “Lithodora II” there are two options: A complete taxo-
nomical conflagration, i.e., reduction of the genera named
under a very widely defined genus Lithospermum, or the
segregation of “Lithodora II” into a new, distinct genus.
Lithodora II” has a variety of morphological characters
clearly distinguishing it from Lithospermum and the other
taxa of this clade (such as shrubby growth form, lack of
faucal appendages and an annulus, and the set of carpo-
logical characters described above) and a segregation of
the group into a new genus seems a more desirable solu-
tion. No published genus name for the group is available
for resurrection, since the only possible genus synonym
of Lithodora is Gymnoleima Decne., which has been typi-
fied on a species of Moltkia (Johnston, 1953b: 27071).
We therefore advocate the creation of a new genus for the
clade “Lithodora II”.
Glandora D.C. Thomas, Weigend & Hilger, gen. nov.
Ty pe : Glandora prostrata (Loisel.) D.C. Thomas
Lithospermum prostratum Loisel. Lithodora pros-
trata (Loisel.) Griseb.
Etymology: The generic name is a blend of glandu-
losus (latinized for glandular, a reference to the glandular
corolla throat) and the genus name Lithodora.
Differt ab Lithodora cicatricibus nucular um appendic-
ulatis canale funiculi inclusis, fauce corollae glandulosa,
corolla extus strigosa vel pubescenta (versus glabra).
Differt ab Lithospermo et Buglossoide absentia annuli
ad basem tubi corollae, fauce corollae non appendiculata,
areolis receptaculi cupulatis vel obliquis; a Buglossoide
etiam differt floribus dimorphis.
Dwarf shrubs with decumbent, ascending, or rarely
erect stems, 560 cm high; underground stolons some-
ti mes present. Leaves alternate, sessile, nar rowly oblong,
narrowly elliptic or narrowly obovate, slightly attenuate
at base, 0.5–6.5 cm long and 1–20 mm wide; leaf margin
entire, sometimes revolute; indumentum hispid, strigose
or tomentose. Inflorescences terminal, 2–10(–14) flow-
ered cymoids (cincinni), frondose-bracteose. Flowers
sessile, subsessile or rarely shortly pedicellate (Glan-
dora oleifolia), actinomorphic, pentamerous, distylous;
calyx 5-cleft almost to base with linear or narrowly
triangular lobes, accrescent at maturity; corolla blue,
purple, or rarely white, narrowly funnel-shaped, abaxi-
ally hairy, adaxially with glandular hairs in the throat,
otherwise glabrous or with 5 patches or a ring of tri-
chomes in the throat or rarely with 5 patches of t richomes
along the veins of the corolla lobes (G. oleifolia) or 5
vertical bands of trichomes extending from the base of
the corolla throat downward to below the middle of the
corolla tube (G. diffusa), 5-lobed, lobes ovate or subcir-
cular, erect, ascending or horizontally spreading, fau-
cal scales absent, annulus absent; stamens 5, included
(long-styled morphs) or slightly exserted (short-styled
morphs), filaments usually inserted at the same height,
rarely at two or three distinctly different levels (G. pros-
trata), filament insertion at the distal part of the corolla
tube (short-styled morphs) or in the median part of the
tube (long-styled morphs), anthers oblong or subellipti-
cal, apex and base rounded or emarginate; gynoeceum
superior, ovary four-parted, with a diffuse nectary in its
proximal portion, gynobase flat, style long and exserted
(long-styled morph) or short and included (short-styled
morph), stigmas 2, terminal at the minutely two-lobed
or notched style apex. Mericarpids usually 1–3(–4) per
flower due to abortion, straight or slightly incurved in
the distal part, ventrally keeled, dorsal side convex, sur-
face smooth or minutely tuberculate (G. prostrata) or
slightly tumulose (G. nitida), whitish, greyish or brown-
ish, cicatrix basal, with one peg-like appendage, which
integrates a sclerenchymatous tubular channel, or rarely
flat apart from a minutely protruding sclerenchymatous
tubular channel in ventral position (G. nitida), areoles
formed by a cupulate depression in the receptaculum,
or rarely areoles oblique with only a weakly developed
depression in ventral position (G. nitida).
Key to the species of Glandora, Lithodora and
allied taxa
1 Corolla tube abaxially hairy and/or corolla throat with
glandular hairs . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1* Corolla tube abaxially completely glabrous, corolla
throat without glandular hairs . . . . . . . . . . . . . . . . 13
2 Stamens inserted just above the base of the corolla.
Flowers small ( < 7 mm), yellow. Corolla lobes glan-
duliferous on both sides . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . Neatostema apulum (L.) I.M. Johnst.
2* Stamen inserted higher up in corolla tube, or when
inserted near the base, then flowers distinctly larger
( > 11 mm) and blue. Corolla lobes not glandulifer-
ous, or only glanduliferous at the base of the adaxial
surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3 Stamens attached at two different levels at the corolla
tube: Two stamina with shorter filaments attached
slightly lower than three stamina with slightly longer
filaments, the two shorter, lower stamina separated
by a longer one. Annulus well developed as a hairy
ring. Calyx circumscissile at the base. Mericarpids
slightly asymmetrical with oblique ventral keel . . .
. . . . . . . . .Moltkiopsis ciliata (Forssk.) I.M. Johnst.
3* Stamens attached at the same level, or when stamina
attached at different levels, then calyx not circumscis-
sile and annulus missing (Glandora prostrata, see 6
and 6*). Mericarpids with straight ventral keel . . 4
4 Areoles flat to slightly concave. Mericarpid append-
age missing or minute. Corolla with annulus and/or
faucal scales or vertical pleats . . . . . . . . . . . . . . . 11
4* Areoles cupulate or rarely oblique with ventral depres-
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
sion. Mericarpids with distinct peg-like appendage or
rarely mericarpid appendage only minute. Annulus
and corolla invaginations missing. Glandora . . . 5
5 Stamina inserted at different levels in the corolla tube.
Mericarpids minutely tuberculate. Glandora pros-
trata (Loisel.) D.C. Thomas . . . . . . . . . . . . . . . . . 6
5* Stamina inserted at one level in the corolla tube. Meri-
carpids smooth or slightly tumulose . . . . . . . . . . 7
6 Prostrate to ascending dwarf shrubs. Adaxial corolla
with a ring of trichomes in the upper corolla tube.
Indumentum of abaxial leaf surface monomorphic
with a continuum of shorter to longer, appressed or
semi-erect, acroscopic trichomes . . . . . . . . . . . . . . .
. .Glandora prostrata (Loisel.) D.C. Thomas subsp.
6* Erect or ascending dwarf shrubs. Adaxial corolla sur-
face with 5 vertical, small fields of trichomes in the
upper corolla throat, or throat nearly glabrous. Indu-
mentum of abaxial leaf surface often dimorphic with
longer, straight, semi-erect to appressed acroscopic
trichomes and much shorter, often bend, appressed
to spreading trichomes . . . . . . . . . . . . . . . . . . . . . . .
. . .Glandora prostrata (Loisel.) D.C. Thomas subsp.
lusitanica (Samp.) D.C. Thomas
7 Leaves adaxially and abaxially with a dense strigose-
tomentose indumentum. Trichomes of both the ad-
and abaxial leaf surface predominantly acroscopic
but some trichomes basiscopic or pointing in various
directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. .Glandora moroccana (I.M. Johnst.) D.C. Thomas
7* Leaves adaxially with a strigose or hispid indumen-
tum. Trichomes, at least of the adaxial leaf surface,
strictly acroscopic . . . . . . . . . . . . . . . . . . . . . . . . 8
8 Leaves obovate, narrowly obovate or narrowly ellip-
tic, not more than four times longer than wide . . 9
8* Leaves narrowly oblong, narrowly elliptic or rarely
narrowly obovate, the larger leaves more than four
times longer than wide . . . . . . . . . . . . . . . . . . . . . 10
9 Leaves silvery because of a dense strigose indumen-
tum on both sides. Mericarpids straight. Flowers very
shortly pedicellate ( –1 mm). Cicatrices almost flat,
apart from a minute, ventrally protruding funicular
channel. Areoles oblique, with only a minor depres-
sion in ventral position . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .Glandora nitida (Ern) D.C. Thomas
9* Adaxial leaf surface with a moderately dense, strigose
indumentum, abaxial leaf surface whitish, with a dense
strigose-tomentose indumentum. Flowers evidently
pedicellate (1–8 mm). Cicatrices with a conspicuous
peg-like appendage. Areoles cupulate . . . . . . . . . . .
. . . . . . . Glandora oleifolia (Lapeyr.) D.C. Thomas
10 Ascending to erect dwarf shrubs. Corolla purple,
adaxially glabrous apart from stipitate glands in the
corolla throat. Leaves with a strigose indumentum of
tightly appressed trichomes . . . . . . . . . . . . . . . . . . .
. . . . . Glandora rosmarinifolia (Ten.) D.C. Thomas
10* Prostrate to ascending dwarf shrubs. Corolla blue,
with 5 vertical lines of trichomes in the upper part
of the corolla tube, and stipitate glands in the corolla
th roat. L eaves wit h a hispid i ndument um of appressed
to semi-erect trichomes . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . Glandora diffusa (Lag.) D.C. Thomas
11 Corolla blue or purple, 15–20 mm long. Corolla throat
and tube exhibiting 5 distinct vertical pleats bearing
hairs and/or stipitate glands, and 5 congregations of
glandular hairs directly beneath the attachment of the
anthers. Perennial . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . Buglossoides Moench sect. Margarospermum
(Rchb.) I.M. Johnst.
11* Corolla yellow, orange, whitish or when blue or purple
corolla either much shorter (–10 mm) or corolla in-
vaginations completely lacking. Corolla invaginations
lacking, or 5 faucal scales present, or when 5 weakly
developed vertical invaginations present, then plants
annual to biennial, corolla at the most 10 mm long and
congregations of glandular hairs directly beneath the
attachment of the anthers missing . . . . . . . . . . . . 12
12 Plants annual or biennial. Corolla inside bearing
5 vertical lines of simple or glandular hairs. Meri-
carpids tuberculate . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . Buglossoides Moench sect. Buglossoides
12* Herbaceous perennials, rarely subshrubs, shrubs or
biennials to annuals. Corolla inside frequently bear-
ing 5 faucal scales, or invaginations lacking. Meri-
carpids usually smooth, or smooth with few to numer-
ous punctate depressions, rarely tuberculate, rugose
or tumulose. . . . . . . . . . . . . . . . . . . . . Lithospermum
13 Corolla distinctly zygomorphic, yellow. Corolla throat
with 5 non-invaginate, rounded to triangular append-
ages borne at or slightly below the level of the filament
attachment . . . . . Halacsya sendtneri (Boiss.) Dörfl.
13* Corolla actinomorphic or with only a very weak ten-
dency to zygomorphy, blue, purplish, pinkish or rarely
white. Corolla throat without appendages . . . . . . 14
14 Annuals. Corolla very small ( < 7 mm in length).
Annulus formed by a villose ring. Calyx at anthesis
tubular for ca. 1/3 of its length, greatly accrescent
at maturity and then tubular for ca. 1/2 to 2/3 of its
length . . .
Mairetis microsperma (Boiss.) I.M. Johnst.
14* Subshrubs or dwarf shrubs. Corolla larger ( > 10 mm
in length). Annulus absent. Sepals free nearly to the
base, calyx at most very shortly tubular . . . . . . . 15
15 Subshrub, producing simple, erect stems from a thick
rhizome. Mericarpid cicatrix without a distinct ap-
pendage, only exhibiting two minute protrusions.
Areoles almost flat . . . . . . . . . . . . . . . . . . . . . . . . . .
. Paramoltkia doerfleri (Wettst.) Greuter & Burdet
15* Much-branched dwarf shrubs or shrubs, rhizomes
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
absent. Mericarpids with one distinct pyramidal to
asymmetrical cicatrix appendage. Areoles cupulate.
Lithodora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
16 Leaves narrowly obovate. Mericarpids muriculate.
The branches which are terminated by an inflores-
cence persisting after the flowering period as leaf-
less shoot thorns. Lithodora hispidula (Sibth. & Sm.)
Griseb. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
16* Leaves linear, narrowly oblong or narrowly ellipti-
cal. Mericarpids rugulose (macroscopically smooth).
Shoot thorns absent . . . . . . . . . . . . . . . . . . . . . . . 19
17 Dwarf shrub, 10–35 cm tall. Cincinni reduced, flow-
ers mostly solitary (1[–4]flowered cincinni) . . . . . .
. . Lithodora hispidula (Sibth. & Sm.) Griseb. subsp.
17* Shrubs, up to 100 cm or more. Inf lorescences: cincinni
or condensed double cincinni to condensed, depau-
perate thyrsoids, (1–)3–6(–more)-flowered . . . . . 18
18 Corolla dark blue. Inflorescences 4–6(–more)-flow-
ered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . Lithodora hispidula (Sibth. & Sm.) Griseb. subsp.
cyrenaica (Pamp.) Brullo & Furnari
18* Corolla with conspicuous ontogenetic colour tran-
sition from white to pinkish to purple, plants often
exhibiting flowers of different colours at once on the
same plant. Inflorescences (1–)3(–5)-flowered . . . .
. . .Lithodora hispidula (Sibth. & Sm.) Griseb. subsp.
versicolor Meikle
19 Corolla purple or blue. Leaves 6–25 mm long. Meri-
carpids straight or slightly incurved, suprabasally
constricted, with a thicker collar at the base. Flowers
exhibiting a stigma height dimorphism . . . . . . . . . .
. . . . . . . . . . . . . . . .Lithodora fruticosa (L.) Griseb.
19* Corolla light blue or rarely white. Leaves 20–40 mm
long. Mericarpids strongly incurved, suprabasally
constricted, with a short, stout neck. Flowers disty-
lous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . Lithodora zahnii (Heldr. ex Halácsy) I.M. Johnst.
Glandora diffusa (Lag.) D.C. Thomas, comb. nov.
Lithospermum diffusum Lag., Varied. Ci. 4: 39. 1805
Lithodora diffusa (Lag.) I.M. Joh nst. in Cont r. Gray
Herb. nov. ser. 73: 56. 1924 – Lectotype (designated
by Valdés, 1981: 1337): [Spain, Asturia], Arvas, VII-
VIII, Lagasca s.n. (MA).
Glandora moroccana (I.M. Johnst.) D.C. Thomas, comb.
nov. Lithodora moroccana I.M. Johnst. in J. Arnold
Arbor. 34: 264. 1953 – Type: [Morocco, Eastern re-
gion], Monte Bu-Ibdiren, Mauricio & Sennen 7946
= Lithospermum diffusum Lag. var. micranthum Faure
& Maire. in Bull. Soc. Hist. Nat. Afrique N. 22: 56.
1931 Lithodora diffusa (Lag.) I.M. Johnst. var.
micrantha (Faure & Maire) Stroh in Beih. Bot. Cen-
tralbl. 58: 212. 1938 – Type: not seen.
Glandora nitida (Ern) D.C. Thomas, comb. nov. Litho-
spermum oleifolium Lapeyr. subsp. nitidum Ern in
Senckenberg. Biol. 49: 79. 1968 Lithodora nitida
(Ern) R. Fern. in Bot. J. Linn. Soc. 64: 73. 1971 – Type:
[Spain, Andalusia], Sierra de Mágina, Provinz Jaén,
Spanien, 1900 m Höhe, 6.6.1963, Ern s.n. (FR!).
Glandora oleifolia (Lapeyr.) D.C. Thomas, comb. nov.
Lithospermum oleaefolium Lapeyr., Hist. Pl. Pyrénées,
Suppl.: 28. 1818 Lithodora oleifolia (Lapeyr.) Griseb.,
Spicil. Fl. Rumel. 2: 531. 1846 – Type: not seen.
Glandora prostrata (Loisel.) D.C. Thomas, comb. nov.
Lithospermum prostratum Loisel., Fl. Gall. 1: 105.
180 6 Lithodora prostrata (Loisel.) Griseb., Spicil.
Fl. Rumel. 2: 85. 1846 – Lectotype (designated here):
[France, Aquitania], sur le bord des chemins et dans
les landes aux environs de Bayonne, 3 juin 1803,
Loiseleur-Deslongchamps s.n. (AV [digital photo-
graph!; “lectotype” noted by Aizpuru & Catalan on
the herbarium sheet, dated 16. 12. 1988]; isotype: AV
[digital photograph!]).
Glandora prostrata (Loisel.) D.C. Thomas subsp. lusi-
tanica (Samp.) D.C. Thomas, comb. nov. Litho-
spermum lusitanicum Samp., Herb. Port.: 123. 1913
Lithodora diffusa (Lag.) I.M. Johnst. subsp. lusitanica
(Samp.) P. Silva & Rozeira in Agron. Lusit. 24: 170.
1962 Lithodora prostrata subsp. lusitanica (Samp.)
Vals in Bol. Soc. Brot. ser. 2. 53.2: 1336. 1981
Lithodora lusitanica (Samp.) Holub in Preslia 58: 302.
1986 – Type: not seen.
= Lithospermum prostratum Loisel. var. erectum Coss.,
Notes Pl. Crit. 1: 42. 1849 Lithospermum diffusum
Lag. var. erectum (Coss.) Rouy, Fl. France 10: 314.
1908 Lithospermum fruticosum L. subsp. diffusum
var. erectum (Coss.) Rouy ex Jahand. & Maire, Cat.
Pl. Maroc. 3: 602. 1934 Lithodora diffusa (Lag.) I.M.
Johnst. var. erecta (Coss.) Stroh in Beih. Bot. Centralbl.
58: 212. 1938 – Type: [Spain, Andalusia], Au Picacho
de Alcalá de los Gazules, 20 Avril 1849, E. Bourgeau
345 (P-CO, as “Bourgeau 245” [photocopy!]). Cosson
(1848: 42) cites a collection with the number 345, the
number 245 on a new herbarium label with otherwise
identical data is almost certainly a transcription error.
Glandora rosmarinifolia (Ten.) D.C. Thomas, comb. nov.
Lithospermum rosmarinifolium Ten., Fl. Napol. Vol.
1. Suppl. 2: 66. 1811–15 Lithodora rosmarinifolia
(Ten.) I.M. Johnst in Contr. Gray Herb. nov. ser. 73:
56. 1924 – Type: not seen.
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
Al-Shehbaz, I. 1991. The genera of Boraginaceae in the south-
eastern United States. J. Arnold Arbor. Suppl. 1: 1–169.
Baldwin, B.G. 1992. Phylogenetic utility of the internal tran-
scribed spacers of nuclear ribosomal DNA in plants: an
example from the Compositae. Molec. Phylog. Evol. 1:
Bañares Baudet, A. (ed.). 2003. Atlas y libro rojo de la flora
vascular amenazada de España: taxones prioritarios. Or-
ganismo Autónomo Parques Nacionales, Madrid.
Bigazzi, M. & Selvi, F. 1998. Pollen morphology in the Bor-
agineae (Boraginaceae) in relation to the taxonomy of the
tribe. Pl. Syst. Evol. 213: 121–151.
Böhle, U.-R., Hilger, H.H. & Martin, W. 1996. Island col-
onization and evo lution of the insular woody habit in
Echium L. (Boraginaceae). Proc. Natl. Acad. Sci. U.S.A.
93: 1174011745.
Boissier, E. 1879. Flora Orientalis: Sive enumeratio plan-
tarum in Oriente a Graecia et Aegypto ad India fines
hucusque observatarum, vol. 4.2. H. Georg, Genevae et
Bolòs i Capdevila, O. de, Font i Castell, X. & Bonada, J.
V. 2001. Atlas Corològic de la flora vascular dels Paï-
sos Catalans, vol. 11. Institut d’Estudis Catalans, Bar-
Browicz, K. 1985. The genus Lithodora Griseb. (Boraginaceae)
in the eastern Mediterranean region. Ann. Mus. Goulandris
7: 39– 48.
Buys, M.H. 2006. A morphological cladistic analysis of Lobo-
stemon (Boraginaceae). S. African J. Bot. 72: 383–390.
Candolle, A.L.P.P. de 18 46. Prodromus systematis naturalis
regni vegetabilis, vol. 10. Fortin, Masson et sociorum,
Casper, B.B., Sayigh, L.S. & Lee, S.S. 1988. Demonstration
of cryptic incompatibility in distylous Amsinckia dougla-
siana. Evolution 42: 248–253.
Cosson, E.S.-C. 1848. Notes sur quelques plantes critiques,
rares, ou nouvelles, et additions à la flore des environs
de Paris. V. Masson, Paris.
Cunningham, C.W. 1997. Can three incongruence tests pre-
dict when data should be combined? Molec. Biol. Evol.
14: 733740.
Farris, J.S., Kallersjo, M., Kluge, A.G. & Bult, C. 19 94. Test-
ing significance of incongruence. Cladistics 10: 15–319.
Felsenstein, J. 1985. Confidence limits on phylogen ies: an ap-
proach using the bootstrap. Evolution 39: 783–791.
Fernandes, R. 1970. Notas sobre a flora de Portugal – X. Bol.
Soc. Brot. ser. 2, 44: 109–170.
Fernandes, R. 1971. Flora Europaea: notulae systematicae ad
floram Europaeam spectantes: no. 8. Bot. J. Linn. Soc.
64: 74–75.
Fernandes, R. 1972. Lithodora. Pp. 88–89 in: Tutin, T.G., Hey-
wood, V.H., Burgess, N.A., Moore, D.M., Valentine, D.H.,
Walters, S.M. & Webb, D.A. (eds.), Flora Europaea, vol.
3. Cambridge Univ. Press, Cambridge.
Ganders, F.R. 1979a. The biology of heterostyly. New Zealand
J. Bot. 17: 607– 635.
Ganders, F.R. 1979b. Heterostyly in Lithospermum cobrense
(Boraginaceae). Amer. J. Bot. 66: 746–748.
Gottschling, M., Hilger, H.H., Wolf, M. & Diane, N. 2001.
Secondar y structure of the ITS1 transcript and its applica-
tion in a reconstruction of the phylogeny of Boraginales.
Pl. Biol. (Germany) 3: 629– 636.
Greuter, W. 1981. Med-checklist notulae 3. Willdenowia 11:
Greuter, W., Burdet, H.M. & Long, G. 1984. Med-Checklist:
A Critical Inventory of Vascular Plants of the Circum-
Mediterranean Countries. Pteridophyta, Gymnospermae,
Dicotyledones (Acanthaceae-Cneoraceae). Conservatoire
et Jardin Botaniques de la Ville de Genéve, Genéve.
Grisebach, A. 18 44– 46. Spicilegium Florae rumelicae et
bithynicae exhibens synopsin plantarum, vol. 2. Frideri-
cum Vieweg et filium, Brunsvigae.
Gürke, M. 1897. Borraginaceae. Pp. 71–131 in: Engler, A. &
Prantl, K. (eds.). Die natürlichen Pflanzenfamilien, vol. 4.
Engelmann, Leipzig.
Hepperle, D. 2001. Multicolor Sequence Alignment Editor. In-
stitute of Freshwater Ecology and Inland Fisheries, Neu-
globsow, Germany.
Hilger, H.H. 1985. Ontogenie, Morphologie und systematische
Bedeutung geflügelter und glochidientragender Cynoglos-
seae- und Eritricheae-Früchte (Boraginaceae). Bot. Jahrb.
Syst. 105: 323378.
Hilger, H.H. & Böhle, U.-R. 2000. Pontechium: a new genus
distinct from Echium and Lobostemon (Boraginaceae).
Taxon 49: 737–746.
Hilger, H.H., Hoppe, J.R. & Hofmann, M. 1993. Energy
dispersive X-ray microanalysis of Boraginaceae subfam.
Boraginoideae nutlet surface: Are Si- or Ca-depositions
characters of systematic relevance? Flora 188: 387–398.
Hilger, H.H., Selvi, F., Papini, A. & Bigazzi, M. 2004. Mo-
lecular systematics of Boraginaceae tribe Boragineae
based on ITS1 and trnL sequences, with special reference
to Anchusa s.l. Ann. Bot. (Oxford) 94: 201–212.
Johnston, I.M. 1924. Studies in the Boraginaceae, III: 1. The
Old World genera of the Boraginoideae. Contr. Gray Herb.
73: 42–78.
Johnston, I.M. 1952. Studies in the Boraginaceae, XXIII: A
survey of the genus Lithospermum. J. Arnold Arbor. 33:
Johnston, I.M. 1953a. Studies in the Boraginaceae, XXIV:
A. Three genera segregated from Lithospermum. B. Sup-
plementary notes on Lithospermum. J. Arnold Arbor. 34:
Johnston, I.M. 1953b. Studies in the Boraginaceae, XXV: A
revaluation of some genera of the Lithospermeae. J. Arnold
Arbor. 34: 258–299.
Johnston, I.M. 1954a. Studies in the Boraginaceae, XXVI:
Further revaluations of the genera of the Lithospermeae.
J. Arnold Arbor. 35: 1–81.
Johnston, I.M. 1954b. Studies in the Boraginaceae, XXVII:
Some general observations concerning the Lithospermeae.
J. Arnold Arbor. 35: 158–166.
ngs tröm, E . & Cha se, M.W. 2002. Tribes of Boraginoideae
(Boraginaceae) and placement of Antiphytum, Echiochi-
lon, Orgastemma and Sericostoma: a phylogenetic analysis
based on atpB plastid DNA sequence data. Pl. Syst. Evol.
234: 137–153.
Li, P. & Johnston, M.O. 2001. Comparative floral morpho-
metrics of distyly and homostyly in three evolutionary
lineages of Amsinckia (Boraginaceae). Canad. J. Bot. 79:
Loidi, J., Bascones, J., Ursua, C. & Casas-Flecha, I. 1988.
TAXO N 57 (1) • February 2008: 79–97Thomas & al. • Phylogeny and systematics of Lithodora
Revision de los Matorrales de la Alianza Genistion Occi-
dentales en las Provincias Vascongadas. Doc. Phytosoc.,
n. s., 11: 311–321.
Meikle, R.D. 1983. Additions to the flora of Cyprus. Ann. Mus.
Goulandris 6: 87–94.
Melchior, H. 1964. Boraginaceae. Pp. 431–434 in: Melchior,
H. (ed.), A. Engler’s Syllabus der Pflanzenfamilien, 12th
ed., vol. 2. Borntraeger, Berlin.
Meusel, H. & Kästner, A. 1990. Lebensgeschichte der Gold-
und Silberdistel: Monographie der mediterran-mittel-
europäischen Compositen-Gattung Carlina, vol. 1. Sprin-
ger Verlag, Wien.
Meusel, H., Jäger, E., Rauschert, S. & Weinert, E. 1978.
Vergleichende Chorologie der zentraleuropäischen Flora,
vol. 2. G. Fischer, Jena.
Mort, M.E., Soltis, P.S., Soltis, D.E. & Mabry, M.L. 2000.
Comparison of three methods for estimating internal sup-
port on phylogenetic trees. Syst. Biol. 49: 160–171.
Olesen, J.M. 1979. Floral morphology and pollen flow in the
heterostylous species Pulmonaria obscura Dumort (Bor-
aginaceae). New Phytol. 82: 757–767.
Philipp, M. & Schou, O. 1981. An unusual heteromorphic
incompatibility system: distyly, self incompatibility, pollen
load and fecundity in Anchusa officinalis (Boraginaceae).
New Phytol. 89: 693–703.
Posada, D. & Crandall, K.A. 1998. Modeltest: testing the
model of DNA substitution. Bioinformatics 14: 817–818.
Ralston, B. 1993. Phylogenetic Systematics and the Evolution
of Mating Systems in Lith ospermu m (Boraginac eae). PhD
thesis. Northern Arizona University, Flagstaff, U.S.A.
Richards, A.J. & Mitchell, J. 1990. The control of incom-
patibility in distylous Pulmonaria affinis Jordan (Borag-
inaceae). Bot. J. Linn. Soc. 104: 369–380.
Riedl, H. 1967. Boraginaceae. Pp. 1–281 in: Rechinger, K.H.
(ed .), Flora Iranica, vol. 48. Akademische Druck- und Ver-
lagsanstalt, Graz.
Riedl, H. 1968. Die neue Tribus der Trigonotideae und das
System der Boraginoideae. Österr. Bot. Z. 115: 291–321.
Riedl, H. 1997. Boraginaceae. Pp. 43–144 in: Kalkman, C.
(ed .), Flora Malesiana, vol. 13. Rijksherbarium/Hortus
Botanicus Leiden University, Djakarta & Leiden.
Sauvage, C. & Vindt, J. 1954. Flore du Maroc: analytique,
descriptive et illustrée, vol. II. Editions Internationales,
Schoen, D.J., Johnston, M.O., L’Heureux, A.M. & Marso-
lais, J.V. 1997. Evolutionary history of the mating system
in Amsinckia. Evolution 51: 1090–1099.
Schou, O. & Philipp, M. 1984. An unusual heteromorphic
incompatibility system. 3. Genetic control of distyly and
self incompatibility in Anchusa off icinalis (Boraginaceae).
Theor. Appl. Genet. 68: 139–144.
Seibert, J. 1978. Fruchtanatomische Untersuchungen an Litho-
spermeae. Dissertationes Botanicae, vol. 44. J. Cramer,
Selvi, F. & Bigazzi, M. 2003. Revision of genus Anchusa
(Boraginaceae-Boragineae) in Greece. Bot. J. Linn. Soc.
142: 431454.
Silva, P. da & Rozeira, A. 1962. Notas sombre algunas plantas
landícolas. Agron. Lusit. 24: 163–176.
Sprengler, H. 1919. Die verschiedenen Typen im Korollenbau
von Lithospermum. Österr. Bot. Z. 68: 109–123.
Stevanović, V., Tan, K. & Iatrou, G. 2003. Distribution of the
endemic Balkan f lora on serpentine. I.: Obligate ser pentine
endemics. Pl. Syst. Evol. 242: 149–170.
Swofford, D. L. 2002. PAUP*: Phylogenetic Analysis Using
Parsimony (* and Other Methods), version 4.0b10. Sinauer
Associates, Sunderland, Massachusetts.
Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Univer-
sal primers for amplification of three non-coding regions
of chloroplast DNA. Pl. Molec. Biol. 17: 1105 –1110.
Tan, K. & Iatrou, G. 2001. Endemic Plants of Greece: The
Peloponnese. Gads Forlag, Kobenhavn.
Vald és , B. 1981. Notas sobre Boraginaceae Españolas. 1. Litho-
dora prostrata (Loisel.) Griseb. y L. diffusa (Lag.) I.M.
Johnston. Bol. Soc. Brot. ser. 2. 53: 1331–1340.
Valdés, B., Rejdali, M., Achhal el Kadmiri, A., Jury, S.L.
& Monserrat, J.M. (eds.). 2002. Catalogue des plantes
vasculaires du nord du Maroc, incluant des clés d’identi-
fication, vol. 2. CSIC, Madrid.
Weigend, M., Hoot, S., Gottschling, M. & Ackermann, M.
2004. A preliminary phylogeny of Loasaceae subfam.
Loasoideae (Angiosper mae: Corna les) based on trnL(UAA)
sequence data, with consequences for systematics and his-
torical biogeography. Org. Div. Evol. 4: 73–90.
Appendix. Voucher data and GenBank accession numbers. An asterisk identifies Genbank accessions reported in previ-
ous studies (Böhle & al., 1996; Hilger & al., 2004).
Taxon: voucher number or collection date (herbarium): collection location; Genbank accession numbers: ITS1, trnLUAA
Outgroup species
Tribe Boragineae: Anchusa officinalis L.: H. Hilger 25.06.2000 (BSB): Germany; AY045710*, AY045703*. Borago officinalis
L.: H. Hilger 25.06.2000 (BSB): Germany; AY383283*, AY383245*. Tribe Echiochileae: Echiochilon fruticosum Desf.: H. Hilger
11/94 (BSB): Israel; EU044843, EU044881. Ogastemma pusillum (Coss. & Durieu ex Bonnet & Barratte) Brummitt: M. Bigazzi &
F. Selvi 04.35 (BSB): Tunisia; EU044842, EU044880.
Ingroup species
Tribe Lithospermeae: Alkanna sieberi A. DC.: A. Kagiampaki 2000/10 (BSB): Greece; EU044844, EU044882. Alkanna tuberculata
(Forssk.) Meikle: Hilger 11.1.00 (BSB): culta HB Graz; EU044845, EU044883. Arnebia coerulea Schipcz.: K. Rechinger 33894
(M): Afghanistan; EU044856, EU044894. Arnebia decumbens (Vent.) Coss. & Kralik: D. Podlech 32857 (M): Tunisia; EU044857,
EU044895. Buglossoides arvensis (L.) I.M. Johnst. subsp. arvensis: M. & K. Weigend 2000/991 (BSB, M, MO): Germany; EU044865,
EU044903. Buglossoides incrassata (Guss.) I.M. Johnst.: H. Hilger 17.05.99 (BSB): Greece, Crete; EU044866, EU044904. Bu-
Thomas & al. • Phylogeny and systematics of LithodoraTAXO N 57 (1) • February 2008: 79–97
glossoides purpureocaerulea (L.) I.M. Johnst.: O. Mohr 582 (BSB): Romania, culta HB Berlin-Dahlem; EU044864, EU044902.
Buglossoides tenuiflora (L. f.) I.M. Johnst.: M. Ristow & B. Seitz 12.03.98 (BSB): Israel; EU044867, EU044905. Echium vulgare
L.: U.-R. Böhle s.n. (BSB): Germany; L43310*, L43312*. Echium wildpretii H. Pearson: U.-R. Böhle s.n. (BSB): Spain, Tenerife;
L43314*, L43316*. Halacsya sendtneri (Boiss.) Dörfl.: D. Uzunov & al. 17.05.2003 (BSB): Yugoslavia; EU044847, EU044885.
Lithodora diffusa (Lag.) I.M. Johnst.: H. Scholz & P. Hiepko 947 (B): Spain; EU044862, EU044900. Lithodora fruticosa (L.) Griseb.
E. Zippel 02–26 (BSB): Spain; EU044853, EU044891. Lithodora hispidula (Sibth. & Sm.) Griseb. subsp. hispidula: A. Kagiampaki
2000/9 (BSB): Greece, Crete; EU044852, EU044890. Lithodora moroccana I.M. Johnst.: D. Podlech 51478 (MSB): Morocco;
EU044860, EU044898. Lithodora nitida (Ern) R. Fern.: D. Thomas 1–05 (BSB): Spain, culta HB Berlin-Dahlem; EU044859,
EU044897. Lithodora oleifolia (Lapeyr.) Griseb.: D. Thomas 9–05 (BSB): Spain, culta HB Berlin-Dahlem; EU044858, EU044896.
Lithodora prostrata (Loisel.) Griseb. subsp. prostrata: P. Garin 17735 (B): Spain; EU044863, EU044901. Lithodora rosmarinifolia
(Ten.) I.M. Johnst.: E. di Gristina 15.02.03 (BSB): Italy; EU044861, EU044899. Lithodora zahnii (Heldr. ex Halacsy) I.M. Johnst.:
D. Thomas 2–05 (BSB): Greece, culta HB Marburg; EU044851, EU044889. Lithospermum afromontanum Weim.: M. Richards
25843 (M): Tanzania; EU044873, EU044911. Lithospermum caroliniense (J.F. Gmel.) MacMill.: R. Thomas & S. Thomas 148354
(NY): U.S.A.; EU044876, EU044914. Lithospermum cinereum DC.: O. Leistner 2109 (M): South Africa; EU044874, EU044912.
Lithospermum distichum Ortega: R. Fernández 2529 (NY): Mexico; EU044879, EU044917. Lithospermum gayanum (Wedd.) I.M.
Johnst.: M. Weigend & C. Schwarzer 8060 (BSB, HUT, USM): Peru; EU044878, EU044916. Lithospermum latifolium Michx.: S.
Hill 30330 (NY): USA; EU044887, EU044915. Lithospermum mirabile Small: R. Worthington 10288 (NY): U.S.A.; EU044875,
EU044913. Lithospermum officinale L.: A. Werres & M. Ristow 15.7.98 (BSB): Germany; EU044872, EU044910. Macromeria
longiflora Sessé & Moç. ex D. Don: J. Rzedowski 34254 (NY): Mexico; EU044871, EU044909. Macromeria viridiflora DC.: E.
Lehto & al. 20212 (NY): U.S.A.; EU044870, EU044908. Mairetis microsperma (Boiss.) I.M. Johnst.: D. Podlech 48277 (MSB):
Morocco; EU044849, EU044887. Moltkia petraea (Tratt.) Griseb.: H. Hilger 22.09.1999 (BSB): culta HB Berlin-Dahlem; EU044854,
EU044892. Moltkia suffruticosa (L.) Brand in Koch: F. Selvi 01.28 (BSB): Italy; EU044855, EU044893. Neatostema apulum (L.)
I.M. Johnst.: M. Bigazzi & F. Selvi 01.09 (BSB): Greece; EU044850, EU044888. Onosmodium bejariense A.DC. in DC.: B. Ert-
ter 5317 (NY): U.S.A.; EU044868, EU044906. Onosmodium molle Michx.: R. Thomas & al. 128151 (NY): U.S.A.; EU044869,
EU044907. Paramoltkia doerfleri (Wettst.) Greuter & Burdet: H. Hilger 25.2.2000 (BSB): Albania, culta HB Berlin-Dahlem;
EU044848, EU044886. Podonosma orientalis (L.) Feinbrun: H. Hilger 26/94 (BSB): Israel; EU044846, EU044884.
Appendix. Continued.
... The samples were sputter coated for 3 min with gold (Spi-Supplies, West Chester, Pennsylvania, U.S.A.). Observations were carried out using a Jeol JSM 6360 (2003), Thomas & al. (2008). Taxa with non-vestured pits are indicated in white, taxa with vestured pits are in black, taxa in grey are not studied. ...
... Within the subfamily Boraginoideae, the species with vestured pits mainly belong to the tribe Lithospermeae ( Fig. 1C-D) (Langström & Chase, 2002;Thomas & al., 2008). Vestured pits are observed in all Lithospermeae species studied. ...
... Nevertheless, in addition to understanding the distribution of vestured pits as ecological adaptation to environmental conditions, there are clear phylogenetic influences in the family. For example, many of the non-vestured Cordia species are more tropical in distribution than most of the Lithospermeae species that are characterized by vestured pits (Gottschling & al., 2004;Thomas & al., 2008). ...
The bordered pit structure in tracheary elements of 105 Boraginaceae species is studied using scanning electron microscopy to examine the systematic distribution of vestured pits. Forty–three species out of 16 genera show a uniform presence of this feature throughout their secondary xylem. Most vestures are small, unbranched and associated with the outer pit aperture of bordered intervessel pits. The feature is likely to have originated independently in the distantly related subfamilies Boraginoideae (tribe Lithospermeae) and Ehretioideae. The distribution of vestures in Ehretia agrees with recent molecular phylogenies: (1) species with vestured pits characterise the Ehretia I group (incl. Rotula), and (2) species with non–vestured pits belong to the Ehretia II group (incl. Carmona). The occurrence of vestured pits in Hydrolea provides additional support for excluding this genus from Hydrophylloideae, since Hydrolea is the only species of this subfamily with vestured pits. Functional advantages of vestured pits promoting parallel evolution of this conservative feature are suggested. The hydraulic benefits of poorly developed vestures remain underinvestigated.
... Unlike most other groups of Boraginaceae, Lithospermeae comprise a considerable proportion of "woody" growth forms, i.e., shrubs, subshrubs, and sclerophyllous shrublets in genera such as Lithodora Griseb., Glandora D.C. Thomas, or Onosma (Weigend et al., 2016, Kolarčik et al., 2010Cecchi et al., 2011). A parallel trend can be observed in Lithospermum, in which the repeated and independent evolution of woody perennials from herbaceous ancestors appears to have taken place in the Mediterranean climate of California (Thomas et al., 2008). Notably, the genus Lobostemon Lehm. is yet another group of sclerophyllous shrubs growing in Mediterranean climates of the Cape Floristic Province (Buys, 2011). ...
... These traits have been termed "key innovations" (Schluter, 2000). Overall, Lithospermeae show a range of particular traits that could be interpreted as potential "key innovations", such as fruticose habit, bird pollination, and heterostyly (increasing outcrossing) in Lithospermum (e.g., Thomas et al., 2008;Cohen, 2011Cohen, , 2012Cohen, , 2016a, and edaphic adaptations in Onosma (e.g., Cecchi et al., 2011). Our results, however, are unable to confirm a direct connection between the appearances of any trait considered and increased diversification rates. ...
Studies about the drivers of angiosperm clade diversifications have revealed how the environment continuously alters the species chances to adapt or to go extinct. This process depends on complex interactions between abiotic and biotic factors, conditioned to the geological and tectonic settings, the genetic variability of species and the rate at which speciation occurs. In this study, we aim to elucidate the timing of diversification of the Lithospermeae, the second largest tribe within Boraginaceae, and to identify the possible morphological and ecological characters associated with shifts in diversification rates of the most species-rich clades. Lithospermeae includes ca. 470 species and 26 genera, among which are some of the largest genera of the family such as Onosma (150 spp.), Echium (60 spp.), and Lithospermum (80 spp.). An exhaustive study of the whole clade is not available to date and its evolutionary history and diversification rates are incompletely known. In the present study, we provide the most comprehensive phylogeny of the group so far, sampling 242 species and all 26 genera. We found that crown-groups and diversification rates of Lithospermeae largely date back to the Mid-Miocene, with high diversification rates in the largest genera, though only significantly high in Onosma. Our analysis fails to associate any of the functional or morphological traits considered with significant shifts in diversification rates. The timing of the diversification of the species-rich clades corresponds with Miocene tectonic events and global climate changes increasing aridity across Eurasia and western North America. These results suggest a causal link between known ecological features of Lithospermeae (i.e., pre-adaptation to arid, open habitats, and mineral soils) and their diversification. Future studies should expand the sampling of individual subclades and detailed functional analyses to identify the contribution of adaptations to arid conditions and pollinator shifts.
... There are no comprehensive phylogenetic studies on the genus Alkanna, although several molecular studies have been published [7][8][9][10]. However, the latter reports described studies on the evolutionary history and phylogenetic relationships in the Boraginaceae family, especially on taxa belonging to Lithospermeae and Boragineae [7][8][9][10][11]. ...
... There are no comprehensive phylogenetic studies on the genus Alkanna, although several molecular studies have been published [7][8][9][10]. However, the latter reports described studies on the evolutionary history and phylogenetic relationships in the Boraginaceae family, especially on taxa belonging to Lithospermeae and Boragineae [7][8][9][10][11]. ...
Full-text available
The Balkans endemic species Alkanna primuliflora Griseb., A. stribrnyi Velen., and A. graeca Boiss. & Spruner have limited distribution in the Balkan Peninsula and a large variation in the morphological characteristics. The populations of the three Alkanna species in the Bulgarian flora are small and fragmented. There are no previous reports on the chemical profile or on the embryology of these species. The hypothesis was that the limited distribution of A. primuliflora, A. stribrnyi, and A. graeca was due to their reproductive capacity and genetic diversity. Furthermore, we hypothesized that the three species will contain pyrrolizidine alkaloids (PAs), as other species of the genus Alkanna (Boraginaceae), but they would have differential alkaloids composition. The population genetic structure and differentiation showed a clear distinction between species and revealed average levels of genetic diversity among the natural populations of the three Alkanna species. The embryological investigation observed stability of the processes in the male and female generative spheres and high viability of mature pollen and embryo of the three species. The normal formation of male and female gametophytes without deviations or degenerative processes, and observed levels of genetic diversity between Alkanna individuals are important in maintaining the size and resilience of the Alkanna populations. Eight alkaloids were identified by GC-MS in A. primuliflora and A. graeca and six alkaloids in A. stribrnyi. The main pyrrolizidine alkaloids (PAs) in all investigated species was triangularine. A. primuliflora and A. graeca showed similar chemical composition that comprised 9-angeloylretronecine, 7-tigloylretronecine, 9-tigloylretronecine, triangularicine, dihydroxytriangularine, dihydroxytriangularicine, whereas, in A. stribrnyi 9-tigloylretronecine, triangularicine and dihydroxytriangularicine were not found. This is the first report on the presence of PAs in A. primuliflora, A. stribrnyi and A. graeca. Besides, this is the first report on the embryology of these endemic species. The results contribute to the knowledge of the three endemic Alkanna species and will facilitate policy-making and defining new strategies for their conservation.
... Generic limits of Lithospermum have been controversial in the past. The phylogeny of Lithospermum and its close relatives has been the focus of several recent publications (Thomas & al., 2008;Cohen & Davis, 2009;Weigend & al., 2009). The combination of molecular and morphological approaches reveals largely north hemispheric Lithospermum as closely allied to two Old World lineages, namely the re-surrected Buglossoides Moench and the newly described Glandora D.C. Thomas, Weigend & Johnst., differing mainly in details of corolla morphology and growth habit (Johnston 1954), are derived lineages of Lithospermum (Weigend & al., 2009). ...
... Figure 2 shows the distribution of the ten species in South America. The originally north-temperate Lithospermum (Langström & Chase, 2002;Thomas & al., 2008;Weigend & al., 2009) reaches its southern distribution limits in NW Bolivia. Lithospermum mediale is widespread in northern Central America and has disjunct populations in northern parts of South America. ...
We present a revision of Lithospermum for South America based on extensive field work and herbarium study which leads to the recognition of ten species including five new to science: Lithospermum bolivariensis, L. cuzcoensis, L. leymebambensis and L. rodriguezii from Peru, and L. azuayensis from Ecuador. The South American species group is morphologically heterogeneous and includes a wide range of growth forms, pollen types and corolla shapes. Molecular data indicate that the species from Ecuador, Peru and Bolivia are monophyletic (South American species group) and represent the sister group of Mexican L. distichum. Lithospermum mediale, disjunct between Colombia/Venezuela and Guatemala/Mexico, represents an independent introduction to South America. Diversification of Andean Lithospermum appears to have progressed from South to North, with the basal taxa being present in the high Andes of southern Peru. The bulk of the species is concentrated in the Amotape‐Huancabamba Zone (northern Peru to southern Ecuador), with five neo‐endemic species being restricted to this area. Lithospermum apparently arrived in the Central Andes ca. 5 Ma and subsequently migrated and diversified northwards, possibly parallel to the uplift of the northern Andes.
... Halacsya is a monotypic genus. The geographical isolation and relation of H. sendtneri to other members of tribe Lithospermae and family Boraginaceae is of substantial study interest to plant ecologists and taxonomists (Thomas et al., 2008;Cecchi & Selvi, 2009;Cecchi et al., 2011). Biochemists have extensively studied biological compounds in the species (Nićiforović et al., 2010;Mašković et al., 2012a, b). ...
... (Balfour 1884, p. 82) and (slightly supported) to the representatives of the "Echium clade". There is no close affinity with Podonosma, which belongs to the "Alkanna clade" (Thomas et al. 2008;Cecchi and Selvi 2009;Chacón et al. 2016Chacón et al. , 2019. Rather, a congeneric relationship can be postulated with Choriantha Riedl (Weigend et al. 2016, p. 72) due to the close resemblance of its solely ascribed species, C. popoviana Riedl with O. angustiloba Riedl; molecular evidence is still missing (an attempt of sequencing the single known specimen at K was unsuccessful; Chacón et al. 2016, p. 538, there spelled "Coriantha"). ...
A complete nomenclatural synopsis is provided for Onosma (Boraginaceae) and the allied genera Maharanga and Podonosma, covering each name of new taxa, each new combination and each replacement name communicated since 1753 till October 2021. All the 749 validly published names are listed together with an account of the formal reasons why 136 additional names must be regarded as invalid. Type species are indicated or newly designated for all the infrageneric names. Information is assembled on the types of the 466 names of specific and infraspecific taxa retrieved from a comprehensive survey of the scientific literature and from public databases, with 126 names (27%) lectotypified here. Finally, 42 new combinations are proposed to solve as yet unresolved nomenclatural issues.
... Moreover, O. orientale Boiss. was accepted by some authors (Thomas et al. 2008., Cohen, 2011., sayadi et al., 2017 as a different genus (Podonosma Boiss.) which was previously named by Boissier (1879). Despite the numerous mentioned studies, the phylogenetic and evolutionary lines of Onosma are poorly recognized (Hilger et al., 2004., Weigend et al., 2009, Kolarcik et al., 2010, Mehrabian et al., 2011a, b., Sheidai et al., 2015. ...
Full-text available
In this paper, seven new species of Onosma are described and illustrated. Besides, a diagnostic key is provided for new species. Hence, based on a wide range of the assessments conducted, the mentioned samples are considered as distinct species when compared with the previously introduced species in the region. Thus, it strengthens our justification as to whether they are new species. Our argument for mentioned novelties is significant differences in the most important morphological and micro-morphological characteristics following trichome type, shape and size of cauline and basal leaves, shape and size of calyx, annulus hairs, shape and size of corolla, nutlets morphology, anther length as well connection and their exertion as well pollen type and size compared to their nearest species. Besides, the recent novelties are confirmed by obviously ecological differentiations.
Onosma atrii and O. persica are described and illustrated as new species and O. caucasica Levin is recorded for the first time from northwestern Iran. The two new species are placed in subsect. Asterotricha (Boiss.) Gürke, because of the asterotrichous indumentum (stellate trichomes) of their basal leaves. An extensive review of the literature and comparison with specimens from various European and Asiatic herbaria led to the recognition of these novelties. Our argument for these innovations is based on observation of significant differences in taxonomically important morphological characters including trichome type, shape and size of basal and cauline leaves, shape and size of calyx and corolla, annulus hairs, anther length, connection type and degree of exsertion, and nutlet morphology. The diagnostic morphological characters of closely related species are discussed, and for each new species a detailed description and details of distribution and conservation status are provided. An updated synoptical key to the species known from the Flora Iranica region and Atropatene sub-province is provided.
Full-text available
A LA VENTA en // Este libro es la primera recopilación moderna que recoge todas las plantas vasculares que viven en la Comunidad Autónoma de Cantabria. ATENCIÓN, NO ES UNA GUÍA FOTOGRÁFICA. En la introducción se da una amplia reseña histórico-botánica y se explica detalladamente la organización del catálogo. La segunda parte se dedica a la descripción geográfica y climática, haciendo hincapié en la descripción fitogeográfica de Cantabria. La parte principal del libro corresponde al catálogo florístico, que consta de 2321 táxones (incluye subespecies e híbridos; 2650 si añadimos los alóctonos subespontáneos). Para cada taxon se da, tras el nombre admitido en negrita, un listado de sus sinónimos principales. Se aportan datos de ecología, y para muchos táxones su filiación fitosociológica. Se explica su corología general y su distribución aproximada en Cantabria mediante una sectorización, aportando su rango altitudinal y el grado de abundancia. Se dan un listado de citas previas (incluyendo su presencia en herbarios). En muchas ocasiones se aporta un listado de localidades inéditas, y se especifican las novedades florísticas para Cantabria. En este trabajo también se han recogido con detalle los táxones alóctonos a los que se les asigna una tipología (naturalizado, invasor, subespontáneo/adventicio y posiblemente autóctono). Tras el catálogo general se aporta como novedad en este tipo de trabajos, un listado de plantas cultivadas. Le sigue la habitual síntesis florística tras la que se dedica un espacio al listado de nombres comunes regionales. El último capítulo abarca la flora protegida en los diversos niveles administrativos y flora la amenazada. La obra finaliza con los correspondientes apartados dedicados a la amplia bibliografía consultada y un completo índice de familias, especies, y nombres comunes.
The recently-developed statistical method known as the "bootstrap" can be used to place confidence intervals on phylogenies. It involves resampling points from one's own data, with replacement, to create a series of bootstrap samples of the same size as the original data. Each of these is analyzed, and the variation among the resulting estimates taken to indicate the size of the error involved in making estimates from the original data. In the case of phylogenies, it is argued that the proper method of resampling is to keep all of the original species while sampling characters with replacement, under the assumption that the characters have been independently drawn by the systematist and have evolved independently. Majority-rule consensus trees can be used to construct a phylogeny showing all of the inferred monophyletic groups that occurred in a majority of the bootstrap samples. If a group shows up 95% of the time or more, the evidence for it is taken to be statistically significant. Existing computer programs can be used to analyze different bootstrap samples by using weights on the characters, the weight of a character being how many times it was drawn in bootstrap sampling. When all characters are perfectly compatible, as envisioned by Hennig, bootstrap sampling becomes unnecessary; the bootstrap method would show significant evidence for a group if it is defined by three or more characters.
Based on measurements of floral parts and experiments, Anchusa officinalis L. is shown to have a distylous breeding system. Dimorphism is found in the style length/anther height ratio, in stigma morphology and in pollen and corolla size. Plants are self-incompatible but intra- and intermorph compatible. Long-styled plants outnumbered short-styled in all populations. Stigmas in short-styled flowers receive a larger quantity of pollen than those of long-styled.