The phylogeny, taxonomy of Macaronesian, European and Mediterranean Roccella (Roccellaceae, Arthoniales)

Full text

The phylogeny and taxonomy of Roccella 405
Symb. Bot. Ups. 34:1
The phylogeny and taxonomy of Macaronesian, European
and Mediterranean Roccella (Roccellaceae, Arthoniales)
Anders Tehler, Åsa Dahlkild, Pia Eldenäs and Guido Benno Feige
Tehler, A., Dahlkild, Å., Eldenäs, P. & Feige, G. B. 2004. The phylogeny and taxonomy of
Macaronesian, European and Mediterranean Roccella (Roccellaceae, Arthoniales). – Acta
Univ. Ups. Symb. Bot. Ups. 34:1, 405–428. Uppsala. ISBN 91-554-6025-9.
The taxonomy, nomenclature and phylogeny are given for eleven Roccella species with focus
on eight species from Macaronesia, Europe and the Mediterranean region. One species, Roc-
cella elisabethae Tehler, is newly described. The names Roccella allorgei des Abb., Roccella
canariensis Darb., Roccella decipiens Darb., Roccella teneriffensis Vain., Roccella fuciformis
f. ensiformis Vain., Roccella fuciformis var. maderensis J.Steiner, Roccella canariensis var.
tuberculata J.Steiner and Roccella canariensis var. subphycopsis J.Steiner are lectotypified. A
key to the Macaronesian and European species is presented. A phylogenetic hypothesis of
supported groups is proposed based on data from the ITS rDNA regions. Forty-three se-
quences from nine Roccella species, of which five species were newly sequenced, have been
produced. The Macaronesian species are resolved as monophyletic. Three so-called species
pairs are discussed: one, Roccella canariensis/R. tinctoria, is still treated as two species,
whereas the other two are treated as single species, Roccella allorgei and R. maderensis.
Key words: lichenized fungi, Macaronesia, ITS, species pair.
Anders Tehler, Naturhistoriska riksmuseet, Sektionen för kryptogambotanik, Box 50007, S-
104 05, Stockholm, Sweden. E-mail: anders.tehler@nrm.se
Åsa Dahlkild, Stockholms universitet, Botaniska institutionen, SE-106 91 Stockholm, Sweden.
E-mail: asa.dahlkild@botan.su.se
Pia Eldenäs, Naturhistoriska riksmuseet, Molekylärsystematiska laboratoriet, Box 50007, SE-
104 05 Stockholm, Sweden. E-mail: pia.eldenas@nrm.se
Guido Benno Feige, Universität Duisburg-Essen, Campus Essen, Fachbereich 9, Botanik/
Pflanzenphysiologie, D-45117 Essen, Germany. E-mail: benno.feige@uni-essen.de
Introduction
Roccella is a genus of lichenized fungi in the fam-
ily Roccellaceae (Arthoniales) that currently count
some 40 species (Follmann 2001). The genus was
described by de Candolle (de Lamarck & de Can-
dolle 1805), and the type is Roccella fuciformis
(L.) DC. It is a fruticose genus recognized by its
grey or grey-brownish thallus with flat or terete
branches, often with white medulla; a cortex with
hyphae anticlinally organized usually into a dis-
tinct palisade plectenchyma (Fig. 10); mainly cir-
cular apothecia with a black hypothecium; fusi-
form, transversely 3-septate spores; and filiform
conidia. The most common secondary metabolites
are erythrin and lecanoric acid. All species use
Trentepohlia as photobiont. Roccella grows on
vertical, overhanging rocks and cliffs or on trees
and shrubs. It is primarily distributed along the
marine, arid, subtropical coastal areas with a Med-
iterranean type of climate such as found in the
Mediterranean area, southern California, the cen-
tral Chilean coast and southern Africa. Macaron-
esia is a region wholly or partially dominated by
such a climate. It consists of several archipelagos
in the Atlantic Ocean off the north-western coast
of Africa: the Azores, Madeira, Canary Islands,
and Cape Verde Islands. The north-western part of
Africa and the south-western part of the Iberian
Peninsula may also be included in the Macarone-

406 Anders Tehler et al.
Symb. Bot. Ups. 34:1
sian region. The Roccella species of Macaronesia
are the main subject of this study. They are largely
restricted to that region but for a few species,
which are also found around the coasts of the Med-
iterranean Sea and along the Atlantic coast as far
north as southern England and with a southern out-
post as far south as northern Angola.
The systematic position of Roccella within the
family Roccellaceae has been the subject of sev-
eral studies using both morphological (Tehler
1983, 1990; Grube 1998) and molecular data
(Myllys et al. 1998, Myllys et al. 1999a). Tehler
(1983) presented a hypothesis where Roccella and
Dirina formed a sister group and where the genera
Dendrographa,Roccellaria,Roccellina and Chio-
decton were found successively more distantly re-
lated to Roccella and Dirina. Later Tehler (1990)
proposed an extended phylogeny including repre-
sentatives from many genera of the order Arthoni-
ales but with emphasis on the Roccellaceae. The
sister group relationship of Dirina and Roccella
was corroborated in that study while Roccellina
and Dendrographa were found to be more dis-
tantly related to that sister pair than was indicated
in the previous study (Tehler 1983). Grube (1998)
used a different selection of genera in his review of
the Arthoniales and he found a close relationship
between Roccella and Dendrographa, but then
neither Roccellina nor Dirina were included. In
the Arthoniales phylogeny presented by Tehler
(1990) terminals were represented by single spe-
cies and the question whether the genera were
monophyletic or not, was not addressed.
Until recently (Myllys et al. 1999a) there were
no dispute over the natural status of Roccella, al-
though most of the generic, diagnostic characters
are seemingly plesiomorphic: fruticose habitat,
cortex texture with anticlinally arranged hyphae,
black hypothecium and 3-septate, fusiform spores.
Myllys et al. (1999a) carried out a study of Roccel-
laceae including five Roccella species. They used
evidence from molecular data (ITS) and the results
indicated that the monophyly of Roccella could be
questioned, since Roccella hypomeca nested with
species of Roccellina. The other Roccella species
of that study (R. montagnei, R. decipiens, R. pe-
ruensis and R. canariensis) were recovered in sis-
ter group relationship together with Dirina as was
earlier suggested by Tehler (1983, 1990) and Myl-
lys (1998).
A few classifications and subdivisions of Roc-
cella have been published, most recently by Foll-
mann (2002) but no extensive phylogenetic hy-
potheses of species relationships within the genus
based on scientific, systematic approaches have
been proposed.
Santesson (1944) divided Roccella into two
sections: Acrocarpon with apical and lecideine
apothecia, and Euroccella with lateral and lecanor-
ine-lecideine apothecia. All species of Roccella
were contained in the section Euroccella except
the single species Roccella minima R.Sant., which
was placed in the section Acrocarpon. That species
was recombined to Protoroccella minima by Foll-
mann (1995). Follmann (1969) used two main
groups of species in the key to Mediterranean Roc-
cella. The first group, the so-called “tinctoria-
type”, was characterized by a cortex texture with
hyphae running uniformly perpendicular to the
surface. The second group, the so-called “canar-
iensis-type”, was characterized by a cortex texture
with the hyphae more irregularly oriented in a
somewhat interlacing pattern similar to a tissue
(see Discussion).
Macaronesian Roccella contains three so-called
species pairs, i.e. pairs of taxa with different repro-
ductive strategies, sexual or asexual but which are
otherwise morphologically, anatomically and
chemically indistinguishable. One pair is the well-
known Roccella tinctoria (sterile) and R. canar-
iensis (fertile) (Myllys et al. 1999b). The compo-
nents of the two other pairs in this study are not
nomenclaturally recognized and both the sterile
and fertile counterparts are treated under the
names, Roccella allorgei and R. maderensis re-
spectively (see The species).
In a similar fashion as that of species pairs with
fertile and sterile morphs as mentioned above, ab-
sence or presence of single lichen substances has
been used to distinguish pairs of species, which are
otherwise morphologically and anatomically in-
distinguishable. Whether such chemical entities
are monophyletic or not is often questioned. Tehler
& Källersjö (2001) studied such a case, Parmeli-

The phylogeny and taxonomy of Roccella 407
Symb. Bot. Ups. 34:1
Table 1. Locality and voucher information for the specimens. Species marked with an asterisk are newly
sequenced. All voucher specimens are deposited in the Swedish Museum of Natural History in Stock-
holm (S).
Species Collector Locality GenBank no.
Dirina catalinariae Tehler 7426 California, Santa Rosa Isl AF138817
Dirina catalinariae Tehler 7530 California, San Luis Obispo AF110352
Roccella allorgei* Tehler 8151 Azores, São Miguel AJ634005
Roccella allorgei* Tehler 8152a Azores, São Miguel AJ634006
Roccella allorgei* Tehler 8209 Azores, Flores AJ634007
Roccella allorgei* Tehler 8226 Azores, Terceira AJ634008
Roccella canariensis Myllys 127a Canary Islands, Tenerife AF110357
Roccella canariensis Myllys 132 Canary Islands, Tenerife AF110358
Roccella canariensis Myllys 138 Canary Islands, Lanzarote AF110359
Roccella canariensis Myllys 139 Canary Islands, Lanzarote AF110360
Roccella canariensis* Tehler 8146 Portugal, Estremadura AJ634009
Roccella canariensis* Tehler 8259 Canary Islands, Tenerife AJ634010
Roccella canariensis* Tehler 8285 Canary Islands, La Gomera AJ634011
Roccella canariensis* Tehler 8309 Canary Islands, Tenerife AJ634012
Roccella decipiens Tehler 7608 Mexico, Baja California AF110354
Roccella elisabethae* Tehler 8272 Canary Islands, Tenerife AJ634013
Roccella elisabethae* Tehler 8306 Canary Islands, Tenerife AJ634014
Roccella fuciformis* Tehler 8171 Azores, São Miguel AJ634015
Roccella fuciformis* Tehler 8255 Canary Islands, Tenerife AJ634016
Roccella fuciformis* Tehler 8288 Canary Islands, La Gomera AJ634017
Roccella maderensis* Tehler 8140 Portugal, Estremadura AJ634018
Roccella maderensis* Tehler 8141 Portugal, Estremadura AJ634019
Roccella maderensis* Tehler 8143 Portugal, Estremadura AJ634020
Roccella maderensis* Tehler 8157 Azores, São Miguel AJ634021
Roccella maderensis* Tehler 8175 Azores, São Miguel AJ634022
Roccella maderensis* Tehler 8188 Azores, Corvo AJ634023
Roccella maderensis* Tehler 8199 Azores, Flores AJ634024
Roccella maderensis* Tehler 8223 Azores, Terceira AJ634025
Roccella montagnei Tehler 7710 South Africa, Cape Peninsula AF110356
Roccella montagnei Tehler 7709 South Africa, Cape Peninsula AF110355
Roccella peruensis Tehler 7523 California, San Nicolas Isl. AF110353
Roccella peruensis* Tehler 7227 Mexico, Baja California Sur AJ634026
Roccella phycopsis* Tehler 8148 Portugal, Estremadura AJ634027
Roccella phycopsis* Tehler 8150a Azores, São Miguel AJ634028
Roccella phycopsis* Tehler 8150b Azores, São Miguel AJ634029
Roccella phycopsis* Tehler 8205 Azores, Flores AJ634030
Roccella phycopsis* Tehler 8221 Azores, Terceira AJ634031
Roccella phycopsis* Tehler 8271 Canary Islands, Tenerife AJ634032
Roccella phycopsis* Tehler 8295 Canary Islands, Tenerife AJ634033
Roccella tinctoria Myllys 135 Canary Islands, Lanzarote AF110364
Roccella tinctoria Myllys 136 Canary Islands, Lanzarote AF110365
Roccella tinctoria Myllys 137 Canary Islands, Lanzarote AF110366
Roccella tinctoria* Tehler 8145 Portugal, Estremadura AJ634034
Roccella tinctoria* Tehler 8153 Azores, São Miguel AJ634035
Roccella tinctoria* Tehler 8179 Azores, Flores AJ634036
Roccella tinctoria* Tehler 8192 Azores, Corvo AJ634037
Roccella tinctoria* Tehler 8214 Azores, Faial AJ634038
Roccella tinctoria* Tehler 8224 Azores, Terceira AJ634039
Roccella tinctoria* Tehler 8257 Canary Islands, Tenerife AJ634040
Roccella tinctoria* Tehler 8287 Canary Islands, La Gomera AJ634041

408 Anders Tehler et al.
Symb. Bot. Ups. 34:1
Roccella tinctoria* Tehler 8308 Canary Islands, Tenerife AJ634042
Roccella tuberculata Myllys 123 Canary Islands, Tenerife AF110361
Roccella tuberculata Myllys 125 Canary Islands, Tenerife AF110362
Roccella tuberculata Myllys 128 Canary Islands, Tenerife AF110363
Roccella tuberculata* Tehler 8144 Portugal, Estremadura AJ634043
Roccella tuberculata* Tehler 8167 Azores, São Miguel AJ634044
Roccella tuberculata* Tehler 8225 Azores, Terceira AJ634045
Roccella tuberculata* Tehler 8258 Canary Islands, Tenerife AJ634046
Roccella tuberculata* Tehler 8310 Canary Islands, Tenerife AJ634047
Species Collector Locality GenBank no.
opsis ambigua and Parmeliopsis hyperopta, and
found that the data from ITS-sequences supported
the two "chemical" species. Another type of chem-
ical variation than the mere absence or presence of
substances has also been used to distinguish spe-
cies, namely the location of a specific substance in
the thallus. One such chemical pair in the present
study is Roccella fuciformis and R. maderensis. In
Roccella fuciformis erythrin is restricted to the sor-
alia only, whereas in R. maderensis it is present
also in the cortex.
In this work we primarily wanted to study spe-
cies delimitations and relationships within the ge-
nus Roccella and whether the Macaronesian spe-
cies were likely to represent a monophyletic group
or not. We also wanted to investigate the relation-
ships between species pair components with re-
gard to their reproductive strategies. Finally, we
wanted to see if morphologically similar taxa with
different distribution of lichen substances within
thalli were supported as species or not.
Materials and methods
Terminal taxa and taxon sampling
The study is based primarily on collections from
the Azores, Canary Islands and Estremadura (Por-
tugal) made by A. Tehler in 2000 and 2001 and on
specimens from the following herbaria: B, BM,
Bradford, COLO, E, ESS, G, GLAM, H, HBG, L,
LINN, M, OXF, PC, S, TUR, UPS and W.
All Macaronesian species were sampled for DNA-
sequencing; Roccella allorgei,R. canariensis,R.
elisabethae (newly described here), R. fuciformis,
R. maderensis,R. phycopsis,R. tinctoria, and R.
tuberculata as well as three widespread Roccella
species from outside Macaronesia; R. decipiens
from California, R. peruensis from North and
South America and, R. montagnei from Asia and
Africa. In sampling the Macaronesian species we
tried to achieve a distribution of the samples geo-
graphically distant from each other. Appointed
voucher specimens are deposited in (S).
The following taxa and samples were taken for
HPLC analysis (voucher specimens): Roccella al-
lorgei sorediate specimen (PC lectotype) and fer-
tile specimen (Tehler 8152, S); R. canariensis (PC
lectotype); R. elisabethae (S holotype); R. fuci-
formis (Tehler 8171, S); R. fuciformis var. ensi-
formis (TUR lectotype); R. maderensis sorediate
specimen (W lectotype) and fertile specimen (Teh-
ler 8140, S); R. phycopsis (Tehler 8148, S).
Altogether 59 ITS1 and ITS2 sequences from
12 species were used in phylogenetic analyses (Ta-
ble 1). One species belongs to Dirina (outgroup)
and 11 belong to Roccella (ingroup). Forty-three
ITS sequences from nine Roccella species were
newly produced for this study. Five Roccella spe-
cies were newly sequenced. Additionally, 14 ITS
sequences of three Roccella species inside and
outside the Macaronesian region were taken from
previous studies and included in the analyses.
At the time of publication Myllys et al. (1999b)
assumed that Roccella tuberculata was one spe-
cies with two varieties, R. tuberculata var. tubercu-
lata and R. tuberculata var. vincentina. For the
sake of convenience Myllys at al. (1999b) referred
Table 1 (cont.)

The phylogeny and taxonomy of Roccella 409
Symb. Bot. Ups. 34:1
only to the species name, R. tuberculata, in the text
and discussion of that paper. This was unfortunate
because the further taxonomic studies of the
present paper have shown that the two varieties are
not conspecific. They are two separate species
with several characteristic features to diagnose
them (see also The species below). To complicate
the situation even further, nomenclatural studies
have shown that R. tuberculata var. vincentina, re-
combined to R. vincentina by Follmann (1987)
cannot keep the epithet of the variety since an ear-
lier name, R. tinctoria, takes priority over R. vin-
centina. The correct names for the two species and
which will be used henceforth are: Roccella tinc-
toria DC. (syn. R. tuberculata var. vincentina
Vain., R. vincentina (Vain.) Follmann) and Roc-
cella tuberculataVain. (syn. R. tuberculata var. tu-
berculata). In Myllys et al. (1999b) the terminals
named as R. tuberculata 123,125 and 128 are to be
referred to the correct name Roccella tuberculata.
The terminals named R. tuberculata 135, 136 and
137 are to be referred to the correct name Roccella
tinctoria. It should be clearly understood that the
species pair discussed in Myllys et al. (1999b) re-
fers to the species called Roccella canariensis and
Roccella tinctoria. What is here referred to as Roc-
cella tuberculata is the name of an obligately ster-
ile sorediate species for which fertile individuals
has never been seen.
With regard to the present study, we have spe-
cifically focused on the monophyletic group of
core Roccella species that constitutes the sister
group to the crustose genus Dirina. The choice of
ingroup was based on previous analyses based on
both morphological and molecular data (Tehler
1983, 1990; Myllys et al. 1998; Myllys et al.
1999a), which indicate and corroborate a sister
group relationship of Dirina and Roccella. Myllys
et al. (1999a) showed that Roccella hypomeca and
Roccellina capensis formed a sister group well
outside Dirina and core Roccella thus indicating
the paraphyly of Roccella.
Molecular and chemical methods
DNA extraction. Extractions of DNA were per-
formed with two different kits. Most specimens
were extracted using QIAGEN’s QIAamp DNA
Mini Kit following the manufacturer’s instructions
with slight modifications: the tissue (a small piece
of thallus) was incubated in tissue lysis-buffer ATL
with proteinase K at 56°C overnight and then mac-
erated with a mini-pestle when needed; the DNA
was eluted once with 50 µl because the yield of
DNA was expected to be low. The Mini-Bead-
Beater with 2.5 mm zirconia/silica beads was used
for grinding the tissue, and the volumes of AP1
buffer and RnaseA were therefore increased to 700
µl and 7 µl respectively. The manufacturer’s in-
structions were followed, though the DNA was
eluted with 50 µl buffer (as above).
PCR amplification. As template, the total DNA ex-
traction (including both fungal and algal DNA)
was used. To avoid amplification of the photobiont
at least one primer needs to be specific for fungi.
We used the PCR-primer pair ITS1F and ITS4 (Ta-
ble 2), where ITS1F is specific for fungi. Amplifi-
cation was performed using Pharmacia Biotech’s
Ready-To-Go™ PCR Beads, according to the
manufacturer’s instructions. The cycling profile
was 95°C 1 min., 45°C 1 min. and 72°C 1 min.,
repeated 40 times. The products were purified with
QIAGEN’s QIAquick PCR purification Kit fol-
Table 2. Primers used in this study. S = Used in sequencing. P = Used in PCR.
Primer Location Sequences Use References
ITS1F in 3' end of 18S 5´-cttggtcatttagaggaagtaa-3´ P Gardes & Bruns 1993
ITS4 in 5' end of 28S 5´-tcctccgcttattgatatgc-3´ P/S White et al. 1990
ITS1LM in 3' end of 18S 5´-gaacctgcggaaggatcatt-3´ S Myllys et al. 1999b
5.8FF in 5.8S 5´-cgcatcttgcgccctccggtat-3´ S this study
5.8FR in 5.8S 5´-cgctcgaacaggcatgcccacc-3´ S this study

410 Anders Tehler et al.
Symb. Bot. Ups. 34:1
lowing the manufacturer’s protocol, except that 44
or 30 µl elution buffer was used and that the eluate
was incubated for 5 min. before last centrifuga-
tion. These modifications were made to increase
sample concentrations.
DNA sequencing. Sequencing reactions were car-
ried out with Applied Biosystems BigDye™ Ter-
minator Cycle Sequencing Kit. Reactions were di-
luted to 1/4 of the manufacturer’s recommenda-
tion, with a total volume of 20 µl. The amount of
template varied from 2–6 µl, depending on the
strength of the PCR product. The primers ITS1LM
and ITS4 were used. Roccella elisabethae speci-
mens were found to have a “poly-A region” to-
wards the 3' end of ITS1, which caused sequencing
difficulties. To solve this, two additional primers
had to be used: 5.8FF and 5.8FR. These internal
primers were designed by comparing the 5.8S re-
gion from several newly sequenced Roccella spe-
cies. All primers utilized in this study are listed in
Table 1. The recommended standard cycling pro-
file was used. For removal of unincorporated ter-
minators the QIAGEN DyeEx 96 Kit was used. All
sequences were assembled and edited using the
Staden Package (Staden 1996). The sequences
have been deposited in GenBank and accession
numbers are given in Table 2.
Chemistry. The chemical investigation was carried
out using gradient elution high-performance liquid
chromatography (HPLC) according to Feige et al.
(1993) if not otherwise stated.
Alignment and phylogenetic analysis
The data was aligned with ClustalX (Thompson et
al. 1997), with some manual adjustment. Gap
opening penalty and gap extension penalty was set
to the default values 15 and 6.66 respectively, in
both the pairwise and multiple alignments.
Phylogenetic analyses were performed with the
parsimony-jackknifing program Xac (Farris
1997), which provides branch swapping and ran-
dom addition sequences. In parsimony jackknifing
the data are internally resampled with a jackknif-
ing technique (Farris et al. 1996) to find well-sup-
ported groups. Resampling works by calculating a
tree for each of a large number of sub-samples
(pseudoreplicates) of characters from the data,
then finding a summary tree, which comprises the
groups occurring in a majority of the trees for sub-
samples. The tree for each pseudoreplicate is
found by parsimony analysis, and each pseudore-
plicate is formed by randomly selecting characters
from the data without replacement, each character
having a fixed chance 1/e (about 37%) of being
excluded. With this resampling technique the ac-
tual number of characters used may vary from rep-
licate to replicate. The resampling procedure can
be repeated up to 10 000 times. The program au-
tomatically discards groups found in less than
50% of the trees for pseudoreplicates, thus elimi-
nating unjustified (poorly supported) resolution
caused by ambiguous data sets. In this analysis the
following settings were used: 1000 pseudorepli-
cates with branch-swapping and 5 random addi-
tion sequences each. Gaps were used as missing
data. Dirina catalinariae Hasse,was used to root
the tree (for details on that species, see Tehler
1983, 2002a). The choice of outgroup species was
based on results from previous analyses (Tehler
1983, 1990; Myllys et al. 1998; Myllys et al.
1999a).
The species
The following account also includes details on the
three Roccella species from outside the Macaron-
esian region to allow and facilitate comparison
with the ingroup, Macaronesian species.
Roccella allorgei des Abb.
Rev. Bryol. Lichénol. 16: 106 (1947). – Type: Portugal,
Azores, Ile de Flores, Sta Cruz, rochers maritimes, 1937,
V. et P. Allorge s.n. (PC lectotype, selected here, PC iso-
lectotypes).
Illustrations. Figs 1–2.
Thallus fruticose; branches usually more or less
terete, or somewhat flattened, with a smooth sur-
face; cortex with palisade plectenchyma; epicor-

The phylogeny and taxonomy of Roccella 411
Symb. Bot. Ups. 34:1
tex gel absent; medulla in holdfast white but usu-
ally streaked with brown or black; soralia com-
mon (sterile morph). Apothecia common, im-
mersed or subimmersed (fertile morph). Spores
21–25 × 6–7 µm.
Chemistry. Thallus and soralia C+ red, branch
tips and soralia K+ pinkish/reddish; contains
erythrin, lecanoric acid, montagnetol and, orsell-
inic acid (trace).
Figure 1. Roccella allorgei,
sorediate specimen. Azores,
São Miguel (Tehler 8151, S).
DNA voucher specimen. Bar
in mm.
Figure 2. Roccella allorgei,
fertile specimen. Azores, São
Miguel (Tehler 8152a, S).
DNA and HPLC voucher
specimen. Bar in mm.

412 Anders Tehler et al.
Symb. Bot. Ups. 34:1
Distribution and habitat.Azores, Madeira, Es-
tremadura (Portugal). On rocks and cliffs.
Remarks.Sorediate morphs of Roccella allorgei
may be morphologically very similar to sorediate
morphs of R. maderensis and can be difficult to
distinguish since both share the same spot test re-
action with C and K. Generally, Roccella allorgei
has more terete branches (Fig. 1) whereas R. mad-
erensis generally has more flattened branches (Fig.
7). Fertile R. allorgei has immersed or subim-
mersed apothecia as distinguished from the sessile
apothecia of R. maderensis.
Roccella allorgei consists of a typical so-called
species pair with a nearly obligately sorediate
morph and a nearly obligately fertile morph. Inter-
mediate specimens with both apothecia and soralia
are uncommon. Roccella allorgei was described
by des Abbayes (1947) based on a sorediate spec-
imen and described as a sorediate taxon. The fer-
tile morph was known and determined by des Ab-
bayes as Roccella africana Vain. Thus, if applying
a species pair terminology to these taxa they would
have been called Roccella allorgei/R. africana.
However, after examining the type material of
Roccella africana from Angola it is evident that it
is not conspecific with the fertile morph of Roc-
cella allorgei. The fertile morph will not be de-
scribed here since the results of this study (see be-
low) clearly indicate that the two morphs are con-
specific and the correct name for the species is
Roccella allorgei.
Roccella canariensis Darb.
Ber. Deutsch. Bot. Ges. 15: 5 (1897) nom.nud.; Ber.
Deutsch. Bot. Ges. 16: 11 (1898b) nom. nud.; Biblioth.
Bot. 45: 50 (1898a) nom. valid. – Roccella tinctoria f.
valida Hampe ex Darbishire, Ber. Deutsch. Bot. Ges. 15:
5 (1897), pro. syn.; Ber. Deutsch. Bot. Ges. 16: 11
(1898b) pro. syn.; Biblioth. Bot. 45: 50 (1898a) pro. syn.
– Type: The Canary Islands "Roccella tinctoria forma
valida, Ins. Canariens, ex herb. Hampe 1877" (G lecto-
type, selected here, G isolectotype).
Roccella hypomecha var. isabellina Vain., in Hiern et
al., Cat. Afr. Pl. (2): 432 (1901). – Type: Angola, Mossa-
medes, “ad rupes maritimas inter Mossamedes et Porto
Pinda loco dicto Cazimba, parce, 1859-09”, Welwitsch
404 (BM lectotype, selected here, TUR isolectotype).
Illustration. Fig. 3.
Thallus fruticose, stiff, brown to greyish-brown;
branches terete, foveate or smooth; cortex hyphae
slightly interlaced, anticlinally arranged but not
forming a conspicuous palisade plectenchyma; ep-
icortex present as a surface gel; medulla in hold-
Figure 3. Roccella canarien-
sis. Canary Islands, Tenerife
(Tehler 8309, S). DNA
voucher collection. Bar in
mm.

The phylogeny and taxonomy of Roccella 413
Symb. Bot. Ups. 34:1
fast white but usually streaked with brown or
black; soralia absent. Apothecia present and usu-
ally numerous, sessile. Spores 21–27 × 5–6 µm.
Chemistry.Thallus C– and K–, cortex C+ red un-
der epicortical layer, disc C+ red; contains erythrin
and lecanoric acid.
Distribution and habitat.Macaronesia and the
western Atlantic coasts of southern Europe and
Africa down to Angola. On rocks and cliffs.
Remarks. Roccella canariensis is the fertile coun-
terpart in a so-called species pair relationship with
the sorediate Roccella tinctoria. Confusingly, Dar-
bishire (1897, 1898a, 1898b) in his publications
cited specimens and their label names as if they
were synonyms. Thus, Darbishire’s lists of syno-
nyms are actually lists of specimens studied. In case
of newly described species the specimens in these
lists should preferably serve as type material and
lectotypes should primarily be chosen among them.
Already when publishing the name Roccella canar-
iensis for the first time Darbishire (1897) cited the
specimen "Roccella tinctoria forma valida
Hampe". The same specimen was later cited when
Darbishire (1898a) validated Roccella canariensis
and therefore it is here selected as lectotype.
Roccella decipiens Darb.
Ber. Deutsch. Bot. Ges. 15: 5 (1897) nom. nud.; Bibli-
oth. Bot. 45: 42 (1898a) nom. valid. – Type: USA, Cali-
fornia, San Diego, 1875, E. Palmer s.n. (G lectotype,
selected here).
Roccella fimbriata Darb., Proc. Calif. Acad. Sci. 21:
287 (1935). – Types: Herb. Calif. Acad. Sci. No. 221700,
collected at South Bay, Cedros Island, Lower California,
Templeton Crocker, August 17, 1932 (not seen). Also
collected at south end of Guadalupe Island, Lower Cali-
fornia, April 1925, H. L. Mason (not seen).
Thallus fruticose, grey; branches flattened,
smooth or rugose; cortex with palisade plecten-
chyma; epicortex gel absent; medulla in holdfast
yellow; soralia absent. Apothecia common, ses-
sile. Spores 22–27 × 5–6 µm.
Chemistry. Thallus and disc C+ red, K+ more or
less pinkish/reddish; contains erythrin, traces of
lecanoric acid (Viethen & Feige 1986 by HPLC,
annotated voucher specimens [M-0024568, M-
0024569]).
Distribution and habitat.On trees, rocks, and
cliffs along the American coast from California in
the north to Peru in the south including adjacent
Pacific islands as well as much of the Caribbean
islands.
Remarks. Roccella decipiens is the fertile coun-
terpart in a so-called species pair relationship with
the sorediate Roccella peruensis (see below). For
details on nomenclature and taxonomy, see Tehler
(2002c).
Roccella elisabethae Tehler sp. nov.
Thallus fruticosus, sterilis, cinereo-albidus; rami numer-
osi, teretes vel subcompressi, 2–10 cm longi. Cortex
hyphis submixtis vel verticalibus. Gelatinum epicorti-
cale non visum. Ascocarpia non visa. Soralia puncti-
formia vel maculiformia. Hypomedulla albida vel brun-
neola. Soralia et thallus C+ rubescens, K non reagens.
Thallus acidum lecanoricum continens.
Type: Canary Islands, Tenerife, Buenavista del Norte,
on the N-slope of Montaña de Talavera near La Cuesta,
28°21.435'N, 16°50.625'W, 300–350 m, 1 Jan. 2001,
Tehler 8306 (S holotype; B, BM, E, ESS, H, LD, NY,
TFMC, UPS, US, isotypes)
Etymology. "elisabethae" refers to the name of the first
authors wife who also took part in the excursion to Ten-
erife in 2001 when the species was collected.
Illustrations. Figs 4–5.
Thallus fruticose; branches terete or slightly flat-
tened, surface smooth or slightly foveate, pruina
present but thin; colour creamy-white; cortex hy-
phae slightly interlaced, anticlinally arranged but
not forming a conspicuous palisade plectenchyma,
smooth, hyaline in lower cortex and brown in up-
per cortex; epicortex gel absent; medulla byssoid,
colour in branches white or sometimes brownish
in various parts; holdfast prominent; hypomedulla
colour (in holdfast) white but streaked with black
or dark brown; soralia present, maculiform or
punctiform; C+ red; K–; isidia absent. Apothecia
absent. Pycnidia absent.

414 Anders Tehler et al.
Symb. Bot. Ups. 34:1
Chemistry. Thallus surface C+ red, K–; medulla
C–, K–; contains lecanoric acid.
Distribution and habitat.Roccella elisabethae
has been collected on the island Tenerife in the
Canary Islands, on Madeira and Porto Santo in the
Madeira Islands and on Santo Antao in the Cape
Verde Islands. It grows on coastal, vertical or over-
hang rocks and cliffs.
Remarks. Roccella elisabethae is habitually more
similar to the sympatric species R. phycopsis than
it is to its closest relatives R. canariensis,R. tinc-
toria and R. tuberculata. It is easily distinguished
from R. phycopsis by the absence of a yellow hy-
pomedulla in the holdfast and its C+ red soralia.
Roccella elisabethae is also similar to the allopat-
ric species R. allorgei but differs by the softer thal-
lus consistency and the K– reaction in its branch
tips and soralia. Roccella allorgei yields a K+ pink
or reddish reaction in the soralia and the outermost
branch tips. Roccella elisabethae has a cortex tex-
ture of slightly interlaced, anticlinally arranged
hyphae (Fig. 5) but it does not form a conspicuous
palisade plectenchyma. A similar type of cortex
structure is shared by R. canariensis,R. tinctoria
and R. tuberculata, but R. elisabethae does not de-
velop the characteristic translucent epicortic gel
that sclerotinizes the upper part of the cortex of
those species (see R. tuberculata, Fig. 13). Roc-
cella elisabethae is a rarely collected species. Oc-
casionally it may be found in some herbaria under
the name Roccella arnoldii Vainio (1901) and at
first we suspected that R. arnoldii would be the
correct name for the species. In the protologue to
Roccella arnoldii Vainio (1901) cited "R. tinctoria
Darb. pr. p. (specimina sorediosa)" and the type
was referred to Arnold’s Lichenes exsiccati no.
1689. The material in Arnold’s exsiccate is all
sorediate and most of it belong to the species Roc-
cella tinctoria. One envelope in (S) also contains
Roccella phycopsis but we have seen no envelopes
containing any material of what is here newly de-
Figure 4. Roccella elisa-
bethae. Canary Islands, Ten-
erife (Tehler 8306, S holo-
type). DNA voucher speci-
men to the right. Bar in mm.
Figure 5. Roccella elisabethae, cortex plectenchyma.
Canary Islands, Tenerife (Tehler 8292, S). Bar 20 µm.

The phylogeny and taxonomy of Roccella 415
Symb. Bot. Ups. 34:1
scribed as R. elisabethae. Even though Vainio
specifies that Roccella arnoldii concerns only part
of Darbishire’s conception of R. tinctoria the cita-
tion of R. tinctoria is here nevertheless interpreted
as an indirect reference to the basionym R. tincto-
ria DC. (Tehler 2002b) since Arnold’s exsiccate
contains nearly exclusively R. tinctoria, which is a
sorediate species. Thus, Roccella arnoldii Vainio
is simply regarded as a superfluous name for Roc-
cella tinctoria. An alternative interpretation would
merely result in Roccella arnoldii being treated as
a taxonomical synonym of R. phycopsis. In no way
can the name R. arnoldii be used for the species
newly described here.
Additional specimens examined.Cape Verde. Santo
Antao, Ribeira da Torre, 17°08'N, 25°04'W, 240 m,
1988, Mies 1142, 9 (G). Portugal.Madeira: Funchal,
Ribeira de St. Luzia, c. 300–450 m, 22 May 1952, Een &
Persson (S); Porto Santo, Persson L52c (S). Spain.Ten-
erife: Adeje, c. 7 km N of Los Cristianos, in Barranco
del Inferno, rocks along trail, 28°07.970'N,
16°42.635'W, 445 m, 2000, Tehler 8267 (S); Anaga Pe-
ninsula, Chamorga, c. 5 km N Igueste, on rock at end of
road in the village, 28°33.787'N, 16°09.250'W, 550 m,
2000, Tehler 8268 (S); Anaga Peninsula, Roque Enme-
dio along road to Taganana, c. 6 km N San Andrés,
28°33.509'N, 16°12.361'W, 450 m, 2000, Tehler 8272
(S); along road S of El Tanque where road crosses old
lava stream El Chinyero, on lava above the road,
28°21.549'N, 16°46.198'W, 520–600 m, 2000, Tehler
8292 (S); San Juan del reparo, Mirador an dessus de Ga-
rachico, 400 m, 15 Mar. 1981, s. coll., s. n. (G); Ladera
de Guimar au dessus la Mirador Don Martia, 150 m, 7
Mar. 1981, s. coll., s. n. (G).
Roccella fuciformis (L.) DC.
Fl. Franç., ed. 3(2): 335 (1805). – Lichen fuciformis L.
Sp. Plant.: 1147 (1753). – Parmelia fuciformis (L.) Ach.,
Meth. Lich.: 258 (1803). – Roccella tinctoria ßfuci-
formis Babington, Primit. Flor. Sarnic.: 123 (1839). –
Roccella fuciforme (L.) Gay, Hist. Fis. Polit. Chile: 89
(1854). – Nemaria fuciformis (L.) Navás, Broteria 8: 3
(1909). – Type: [icon in] Dillenius, Hist. musc. T. 22, f.
61A&B, 1742: Lichenoides fuciforme tinctorium, cor-
niculis longioribus & acutioribus (lectotype; OXF epit-
ype [the large, central specimen, second from right]; both
designated and selected by Jørgensen et al. 1994: 314).
Roccella teneriffensis Vain., Kongel. Danske Vidensk.
Selsk. Skr., Naturvidensk. Math. 8(6.3): 397 (1924). –
Type: Canary Islands, Tenerife, Orotava, 5/1 1921, F.
Børgesen 8 (TUR-V 26936 lectotype, selected here).
Roccella fuciformis f. ensiformis Vain., Kongel. Dan-
ske Vidensk. Selsk. Skr., Naturvidensk. Math. 8(6.3):
397 (1924). – Type: Canary Islands, Tenerife, “ad saxa
vulcania in Orotava Ster.” (TUR-V 26987 lectotype, se-
lected here, TUR-V 26988, 26989 isolectotypes).
Illustration. Fig. 6.
Thallus fruticose, grey or sometimes brownish;
branches flattened, smooth; cortex with palisade
plectenchyma; epicortex gel absent; medulla in
holdfast white but usually streaked with brown or
black; soralia common. Apothecia rare, sessile.
Spores 20–23 × 5 µm.
Chemistry. Thallus C– or in parts faintly red-
dish, soralia C+ red, K+ more or less yellowish
to sometimes nearly K–; contains erythrin, lep-
raric acid (also erythrin, lepraric acid by
Sanchez-Pinto 1987 by TLC, annotated voucher
specimen [TUR-V 26987 lectotype for Roccella
fuciformis f. ensiformis Vain.]).
Distribution and habitat.Macaronesia, the
western Mediterranean Region and the Atlantic
coast from Morocco in the south to England in the
north. On rocks and cliffs, rarely on trees.
Remarks. Roccella fuciformis and R. maderensis
are morphologically two confusingly similar spe-
cies, in the field often indistinguishable. Normally,
Roccella fuciformis is C–, K+ yellowish (thallus
surface) and R. maderensis is C+ (thallus surface)
and K+ reddish/pinkish (soralia and branch tips).
However, reactions are sometimes faint or occa-
sionally even contradictive (see also Discussion).
Roccella fuciformis is rarely fertile and when
apothecia are present, soralia are present as well.
This species, as it appears, is solely or predomi-
nantly dependent on soredia for its dispersal. It has
never been found exclusively with apothecia.
Dubious taxa referred to as Roccella fuciformis:
Roccella fuciformis var. arborea DC., Fl.
Franc., ed. 3(2): 335 (1805). – Type: “prés Saint-
Malo, par le C. du Petit-thours; en Provence, par le
C. Deleuze” (not found). This must be either a cor-
ticolous specimen of Roccella fuciformis or R.
phycopsis.

416 Anders Tehler et al.
Symb. Bot. Ups. 34:1
Roccella fuciformis var. ventricosa Mont., Ann.
Sci. Nat. Bot. 7: 145 (1857). – Type: “Hab. Ad
rupes maritimas insularum Grogonearum (Cap
Vert) a cl. Bolle lecta.” (not found). Follmann &
Mies (1988) concludes that the type is lost and
suggests this may be a malformed Roccella fuci-
formis. However, Montagne (1857) described Roc-
cella fuciformis var. ventricosa as a fertile taxon
with no mentioning of soralia. Roccella fuciformis
is extremely rarely, if ever, found solely fertile. Per-
haps it is more likely that Roccella fuciformis var.
ventricosa is actually a malformed R. canariensis.
Roccella fuciformis var. immutata J.Steiner, Ös-
terr. Bot. Z. 54: 408 (1904). – Roccella immutata
(J.Steiner) Follmann, Courier Forschungsinst.
Senckenberg 159: 179 (1993). – Type: “Gran Ca-
naria, prope Tafira, lavicola. Nur in einigen
Büschelchen gesammelt, nach Bormüller: ‘an den
dem NW-Passat ausgesetzten Felswänden ca. 400
m s m’” (not found). Obviously the type material
is very scarse. According to Steiner both cortex
and soralia are C negative. Thus, it seems reasona-
ble to suggest that it may be a small or poorly
developed specimen of Roccella tuberculata, the
only species in Macaronesia showing that type of
C reaction. Follmann’s exsiccate Lich. exs sek,
mus. bot. berlin. 53 Roccella immutata J.Steiner in
(H) is identical to R. tuberculata.
Roccella maderensis (J.Steiner) Follmann
J. Hattori Bot. Lab. 32: 37 (1969), nom invalid.; Courier
Forschungsinst. Senckenberg 159: 180 (1993) nom.
valid. – Roccella fuciformis var. maderensis J.Steiner,
Österr. Bot. Z. 54: 408 (1904). – Nemaria fuciformis var.
maderensis (Steiner) Navás, Broteria 11: 32 (1913). –
Type: Portugal, Madeira, Ponte Delgada (W-806 lecto-
type, selected here [excluding upper specimen to left,
which is = R. fuciformis], PC, W-12989 isolectotypes).
Illustrations. Figs 7–8.
Thallus fruticose, grey to brownish; branches flat-
tened, smooth or slightly rugose; cortex with pal-
isade plectenchyma; epicortex gel absent; me-
dulla in holdfast white but usually streaked with
brown or black; soralia common. Apothecia rare,
sessile. Spores 20–27 × 6 µm.
Chemistry. Thallus C+ red but sometimes reac-
tion may be faint, soralia C+ red, branch tips
(rarely all thallus) and soralia K+ pinkish/reddish;
contains erythrin, lecanoric acid and montagnetol.
Distribution and habitat.Azores, Madeira and
Estremadura (Portugal). On rocks and cliffs.
Remarks. Roccella maderensis and R. fuciformis
are morphologically two confusingly similar spe-
cies, in the field often indistinguishable. Normally
Figure 6. Roccella fuciformis,
Canary Islands, Tenerife
(Tehler 8312, S). Bar in mm.

The phylogeny and taxonomy of Roccella 417
Symb. Bot. Ups. 34:1
R. maderensis is C+ (thallus surface) and K+ red-
dish/pinkish (soralia and branch tips) and Roccella
fuciformis is C–, K+ yellowish (thallus surface).
However, the reactions are sometimes faint or oc-
casionally even contradictive (see also Discus-
sion). Sorediate morphs of R. maderensis may also
be morphologically very similar to sorediate
morphs of R. allorgei, from which it can be diffi-
cult to distinguish since both share the same spot
test reaction with C and K. Generally, Roccella
maderensis has more flattended branches (Fig. 7)
whereas Roccella allorgei generally has more
terete branches (Fig. 1). Fertile Roccella maderen-
sis has sessile apothecia, which are usually easily
distinguished from immersed or subimmersed
apothecia of R. allorgei.Roccella maderensis con-
sists of a typical so-called species pair with a
nearly obligately sorediate morph and a nearly ob-
ligately fertile morph. Intermediate specimens
with both apothecia and soralia are rare. Roccella
maderensis was described by Steiner (1904) based
on a sorediate specimen and described as a soredi-
Figure 7. Roccella maderen-
sis, sorediate specimen.
Azores, Corvo (Tehler 8188,
S). DNA voucher specimen.
Bar in mm.
Figure 8. Roccella maderen-
sis, fertile specimen. Portu-
gal, Estremadura, Cabo da
Roca (Tehler 8140, S). DNA
and HPLC voucher specimen.
Bar in mm.

418 Anders Tehler et al.
Symb. Bot. Ups. 34:1
ate taxon. The fertile morph has never been de-
scribed and if collected it is usually determined as
R. fuciformis. The fertile morph will not be de-
scribed here since the results of this study (see be-
low) clearly indicate that the two morphs are con-
specific and the correct name for the species is
Roccella maderensis.
Roccella montagnei Bel.
Voyage Indes-Orient. 2 Cryptog.: 117 (1838). – Type:
India, Pondecherry, “ad Mangiferas indicas”, s.ann., Be-
langer s.n. (PC lectotype, selected by Awasthi (1981:
218), PC isolectotype).
Thallus fruticose, grey to greenish-grey; branches
flattened, rugose or sometimes smooth; cortex
with palisade plectenchyma; epicortex gel absent;
medulla in holdfast white usually streaked with
brown or black; soralia common. Apothecia ab-
sent or very rare, sessile. Spores not found.
CHEMISTRY. Thallus C+ red and at least branch tips
K+ pinkish/reddish, soralia C–; contains erythrin,
traces of lecanoric acid, roccellic acid, unidenti-
fied substance (Johnson & Culberson 1980, anno-
tated voucher specimen [lectotype of Roccella
montagnei PC]); erythrin, traces of lecanoric acid
and unidentified substance [RT 14.30] (Viethen &
Feige 1986 by HPLC, annotated voucher speci-
mens [BM 000661005, BM 000661010, BM
000660970, BM 000660960, BM 000660961, BM
000660962, BM 000660963, M-0024582, M-
0024579]).
DISTRIBUTION AND HABITAT.Asia, Oceania and Af-
rica. On trees and shrubs.
REMARKS. Roccella montagnei is the sorediate
counterpart in a so-called species pair relationship
with the fertile Roccella belangeriana Awasthi
(Awasthi 1981). Both taxa belong in a large and
critical species complex distributed mainly in Asia
and Africa but not present in the Americas. De-
scribed species of this group are: R. pattensis
C.W.Dodge, R. pembensis C.W.Dodge, R. appla-
nata Choisy, R. endocrocea Choisy, R. intermedia
Choisy, R. linearis auct., R. podocarpa Vain., and
possibly R. sinensis Nyl. In addition to those
names seven taxa of various subspecific ranks have
also been described.
Roccella peruensis (Krempelh.) Darb.
Bull. Herb. Boiss. 5: 763 (1897). – Roccella montagnei
var. peruensis Krempelh., Verh. Zoo.-bot. Ges. 26: 442
(1877). – Type: Peru, Lima, 1868–71, Barranca s.n. (M-
0024563 lectotype, selected by Tehler 2002c).
Thallus fruticose, grey; branches flattened, rugose
or sometimes smooth; cortex with palisade plect-
enchyma; epicortex gel absent; medulla in hold-
fast yellow; soralia common. Apothecia absent.
CHEMISTRY. Thallus C+ red and K+ more or less
pinkish/reddish, soralia C–; contains erythrin, le-
canoric acid (Johnson & Culberson s.ann. by TLC,
annotated voucher specimen [M-0024609, H, S,
COLO-363839], Lichenes selecti exsiccati, Vûzda
1882).
DISTRIBUTION AND HABITAT.The Pacific coast from
North America (California), Central America in-
cluding the Caribbean, to South America (Peru).
On rocks and cliffs.
REMARKS. Roccella peruensis is the sorediate
counterpart in a so-called species pair relationship
with the fertile Roccella decipiens (see above). For
more details on nomenclature and taxonomy, see
Tehler (2002c).
Roccella phycopsis (Ach.) Ach.
Lich. univ.: 440. 1810. – Lichen fucoides Dicks., Pl.
Crypt. Brit. 2: 22 (1790), nom illeg., non Lichen fu-
coides Neck., Meth. Musc.: 76 (1771), nec Lichen fu-
coides With., Bot. Arrang. 1: 726 (1776). – Parmelia
phycopsis Ach., Arch. Syst. Naturgesch. 1(1): 110
(1804). – Roccella fucoides Vain., in Hiern et al., Cat.
Afr. Pl. 2: 433 (1901). – Type: “Habitat in rupibus et
lignis, locis maritimis, prope Gosport”, not found; “Li-
chen fucoides”, s.loc., s.ann., Dickson s.n. (E neotype,
selected by Tehler 2002b, E isoneotype).
Roccella pygmaea Durieu & Mont., Flore d’Alger:
226 (1847). – Roccella pusilla de Notaris, Giorn. Bot.
Ital. 2: 222 (1846). – Roccella phycopsis f. pygmaea
(Durieu & Mont.) Nyl. ex Zahlbr., Catal. Lich. Univ. 2:

The phylogeny and taxonomy of Roccella 419
Symb. Bot. Ups. 34:1
519 (1923). – Roccella fucoides f. pygmaea (Durieu &
Mont.) Zahlbr., Catal. Lich. Univ. 2: 519 (1923). – Type:
“Algeria. ad ficos et Micromeriam inodoram”, s.ann., M.
Durieu s.n. (PC lectotype, selected by Tehler 2002b, PC
isolectotypes.
Roccella phycopsis f. tenuior Leight., Lich. Flor. Brit.
ed 3:74 (1879). – Roccella fucoides f. tenuior (Leight.)
Zahlbr., Catal. Lich. Univ. 2: 519 (1923). – Type: Great
Britain, Channels Islands, Jersey, La Moye, s.ann., Lar-
balestier s.n. (BM lectotype, selected by Tehler 2002b).
Roccella phycopsis var. ceciliametella Rabenh.,
Flecht. Europ. 36: no 958 (1879), nom. nud. – Roccella
fucoides var. ceciliametella (Rabenh.) Zahlbr., Catal.
Lich. Univ. 2: 519 (1923), nom. nud. – Coll. orig.: Italy,
Sur le tombeau de Cecilia Metella, 1874, Vito Beltrani
Pisani in Rabenhorst, Lich. Eur. exs. 958 (PC, S).
Roccella fucoides var. corticola Sambo, Nouvo Giorn.
Bot. Ital. 34: 832 (1927). – Type: [Greece, Rhodes]
“Sugli alberi a Cattavia”, 1926, Senni s.n. (not found).
Roccella fucoides var. farinosa Delise ex Sambo,
Nouvo Giorn. Bot. Ital. 34: 832 (1927). – Type: France,
St Malo, s.ann., s.coll. s.n. (PC lectotype, selected by
Tehler 2002b).
ILLUSTRATIONS. Figs 9–10.
Thallus fruticose, grey to cream coloured;
branches terete, foveaete or nearly smooth; cortex
with palisade plectenchyma; epicortex gel absent;
medulla in holdfast yellow; soralia common.
Apothecia rare, sessile. Spores 25–31 × 5–7 µm.
CHEMISTRY.Thallus C+ red, K– or K+ reddish, sor-
alia C–; contains erythrin, lecanoric acid.
DISTRIBUTION AND HABITAT.Macaronesia, all of the
Mediterranean Region and the Atlantic coast from
Morocco in the south to England in the north. On
rocks and cliffs, rarely corticolous.
Figure 9. Roccella phycopsis.
Portugal, Estremadura, Cabo
da Roca (Tehler 8148, S).
DNA and HPLC voucher
specimen. Bar in mm.
Figure 10. Roccella phycopsis, cortex palisade plecten-
chyma. Portugal, Estremadura, Cabo da Roca (Tehle
r
8148, S). Bar 20 µm.

420 Anders Tehler et al.
Symb. Bot. Ups. 34:1
REMARKS. Roccella phycopsis is rarely fertile and
when apothecia are present, soralia are present as
well. This species, as it appears, is solely or pre-
dominantly dependent on soredia for its dispersal.
It has never been found exclusively with apothe-
cia. For further details on nomenclature and taxon-
omy, see Tehler (2002b, 2003).
Roccella tinctoria DC.
Fl. Franç. ed. 3,2: 334. (1805). – Lichen roccella L., Sp.
Pl.: 1154. (1753). – Parmelia roccella (L.) Ach., Meth-
odus: 274 (1803). – Thamnium roccella (L.) A.St.-Hil.,
Expos. Fam. Nat. 1: 21 (1805). – Roccella purpura-an-
tiquorum Bory, Dictionn. Class. Hist. Nat. 14: 630
(1828). – Roccella tinctoria f. filiformis C.Bab., Primit.
Fl. Sarnic.: 123 (1839). – Roccella arnoldii Vain., in Hi-
ern et al., Cat. Afr. Pl. (2): 433 (1901). – Nemaria roc-
cella (L.) Navás, Broteria 8: 4 (1909). – Roccella fu-
coides var. arnoldii (Vain.) Zahlbr., Catal. Lich. Univ. 2:
519 (1923). – Type: Roccella ex insula Fyal, s.ann.,
s.coll. s.n. (LINN 1273.263, lower specimen lectotype,
LINN 1273.263 upper specimen isolectotype), [lecto-
type designated by Howe (1912); lectotype designation
narrowed (Art. 9.14) by Jørgensen et al. (1994)].
Physcia dichotoma Pers., Ann. Wetterauischen Ges.
Gesammte Naturk., Frankfurt. 1(2): 18 (1810). – Roc-
cella tinctoria var. dichotoma (Pers.) Ach., Syn. Meth.
Lich.: 337 (1814). – Roccella dichotoma (Pers.) Ach.
Syn. Meth. Lich.: 337 (1814), pro syn. – Type: (Coll.
orig. searched for but not found in L); “Roccella di-
chotoma” in herb Persoon, s.loc., s.ann., s.coll. s.n. (L
0246980 neotype, designated by Tehler 2002b, L
0246981 isoneotype).
Roccella fastigiata Bory, Dictionn. Class. Hist. Nat.
14: 631 (1828). – Roccella tinctoria var. fastigiata
(Bory) Delise ex Darb., Biblioth. Bot. 45: 32 (1898a),
pro syn. et comb. nov. inval. – Type: Coll. orig.: [France]
Ouessant, Pylaie; côtes de Saint-Malo, Brébisson
(searched for but not found in PC); Canary Islands, Ten-
erife, 1822, Delise s.n. (PC neotype, designated by Teh-
ler 2002b).
Roccella patellata Stirton, J. Linn. Soc. Bot. 14: 366
(1875). – Type: [Cape Verde Is.], “ad saxa, Bird Island”,
1843, Stirton s.n. (BM-660991! lectotype, selected by
Tehler 2002b; BM-660954, BM-660954, GLAM-STIR-
TON isolectotypes).
Roccella tuberculata var. vincentina Vain., in Hiern et
al., Cat. Afr. Pl. (2): 433 (1901). – Roccella canariensis
var. vincentina (Vain.) Zahlbr., Deutsch. Südpolar-Exp.
1901–1903. 8: 21 (1906). – Roccella vincentina (Vain.)
Follmann in Follmann & Sánchez-Pinto, Courier Forsc-
hungsinst. Senckenberg 95: 184 (1987). [ut “vicentina”]
– Type: Cape Verde Is., “ad rupes maritimas insulae S.
Vicentii, Cap. Virid.”, January 1861, Welwitsch 402
(BM-660953 lower left specimen, lectotype, selected by
Tehler 2002b, isolectotypes: BM-660953, TUR-V
26958).
Roccella boergesenii Vainio, Kongel. Danske Vi-
densk. Selsk. Skr., Naturvidensk. Math. Afd. 8: 397
(1924). – Type: Canary Islands, Tenerife, Orotava
(Taoro), 6 Jan. 1921, Børgesen s.n. (TUR-V 26919 lec-
totype, selected by Tehler 2002b).
Roccella guanchica Feige & Viethen, Herzogia 7: 611
(1987). – Type: Canary Islands, Tenerife, ad saxa vul-
canica in territorio “Malpais de Güimar”, ad mare, circ.
20 m.s.m., expositione septentriale usque ad occiden-
tale, 1985, Feige & Viethen s.n. (ESS 3447 holotype).
ILLUSTRATION. Fig. 11.
Figure 11. Roccella tinctoria.
Canary Islands, Tenerife
(Tehler 8308, S). DNA
voucher specimen. Bar in
mm.

The phylogeny and taxonomy of Roccella 421
Symb. Bot. Ups. 34:1
Thallus fruticose, stiff, brown to greyish-brown;
branches terete, foveaete or nearly smooth; cortex
hyphae slightly interlaced, anticlinally arranged
but not forming a conspicuous palisade plecten-
chyma; epicortex present as a surface gel; me-
dulla in holdfast white but usually streaked with
brown or black; soralia present and usually nu-
merous. Apothecia absent.
CHEMISTRY. Thallus C–, K–, soralia C+ red and
cortex under epicortical layer C+ red; contains
erythrin (sometimes traces or absent), lecanoric
acid (Viethen & Feige 1986 by HPLC and TLC,
annotated voucher specimens [BM 000660953,
ESS 3447, M-0024629, M-0024628, M-0024638,
M-0024695]).
DISTRIBUTION AND HABITAT.Macaronesia and the
western Mediterranean Region, the Atlantic coast
from Portugal south to Cape Verde Islands. On
rocks and cliffs.
REMARKS. Roccella tinctoria is the sorediate coun-
terpart in a so-called species pair relationship with
the fertile Roccella canariensis (see Discussion).
Roccella tuberculata Vain.
Catal. Welwitsch Afric. Plants, 2: 433 (1901). – Roccella
canariensis var. tuberculata (Vain.) J.Steiner Österr.
Bot. Z. 9: 29 (1904). – Type: Cape Verde. "Ad rupes in
insula Fogo leg. Pharmacop. Torres. No. 419. Ster.” (BM
000660995 lectotype, selected here, TUR-V 26957 iso-
lectotype).
Roccella canariensis var. subphycopsis J.Steiner, Ös-
terr. Bot. Z. 9: 30 (1904). – Type: Ponte Delgada, Ma-
deira, 1901, Bornmüller s.n. (W lectotype, selected
here).
ILLUSTRATIONS. Figs 12–13.
Thallus fruticose, stiff, brown to greyish-brown;
branches terete, foveaete or nearly smooth; cortex
hyphae slightly interlaced, anticlinally arranged
but not forming a conspicuous palisade plecten-
chyma; epicortex present as a surface gel; me-
dulla in holdfast dirty yellow; soralia present and
usually numerous. Apothecia absent.
CHEMISTRY. Thallus C–, K–, soralia C–, cortex un-
der epicortical layer C+ red; contains lecanoric
acid (Viethen & Feige 1986 by HPLC, annotated
voucher specimens [BM 000660995 lectotype, M-
0024628]).
Figure 12. Roccella tubercu-
lata. Portugal, Estremadura,
Cabo da Roca (Tehler 8144,
S). DNA voucher specimen.
Bar in mm.

422 Anders Tehler et al.
Symb. Bot. Ups. 34:1
DISTRIBUTION AND HABITAT.Macaronesia and the
Atlantic coast from Portugal south to Cape Verde
Islands. On rocks and cliffs.
REMARKS.Habitually very similar to R. tinctoria,
but distinguished by the negative hypochlorite re-
action in the soralia and the holdfast with a dirty
yellow medulla. Roccella tuberculata does not
have a fertile counterpart. It has never been found
with apothecia and as far known it is obligately
sterile.
Key to the Macaronesian, European and Mediterranean Roccella species
For an overview key to the fruticose, subfruticose and subfoliose genera in Roccellaceae the reader is referred to
Tehler et al. (1997).
1. Cortex surface C+...............................................................................................................................2
– Cortex surface C–...............................................................................................................................7
2. Soralia present, sometimes apothecia also present.............................................................................3
– Soralia absent, apothecia present ......................................................................................................6
3. Soralia and branch tips K+ pinkish, branches often 10 cm or longer ................................................4
– Soralia and branch tips K–, branches usually no longer than 5 cm ...................................................5
4. Branches mostly smooth and slender, flattened or nearly terete .........................................R. allorgei
– Branches mostly rugose and broader, more or less flattened ........................................R. maderensis
5. Soralia C+ red, hypomedulla brown-white in holdfast .................................................R. elisabethae
–Soralia C–, hypomedulla yellow in holdfast ................................................................... R. phycopsis
6. Branches mostly smooth and slender, flattened or nearly terete, apothecia subimmersed...................
.............................................................................................................................................R. allorgei
– Branches mostly rugose and broader, more or less flattened, apothecia sessile............R. maderensis
7. Apothecia present...............................................................................................................................8
– Apothecia absent ................................................................................................................................9
8. Subcortex C–, branches flattened, smooth ......................................................................R. fuciformis
– Subcortex C+ red, branches terete, foveaete.................................................................R. canariensis
9. Subcortex C–, branches flattened, smooth ......................................................................R. fuciformis
– Subcortex C+ red, branches terete, foveaete....................................................................................10
10. Soralia C+ red, hypomedulla brown-white in holdfast ..................................................... R. tinctoria
– Soralia C–, hypomedulla dirty yellow in holdfast.........................................................R. tuberculata
Figure 13. Roccella tuberculata, cortex and gelatinous
epicortex. Canary Islands, Tenerife (Tehler 8310, S). Ba
r
5 µm.

The phylogeny and taxonomy of Roccella 423
Symb. Bot. Ups. 34:1
Results
The sequences varied in length between 547 and
644 base pairs (bp) and most of them include the
end of 18S and complete ITS1, 5.8S, and ITS2 and
the beginning of 26S. For one sample, Roccella
tinctoria 8145, we were unable to sequence about
150 bp in the 5' end of ITS and this sequence is
thus only 456 bp long. The data set used for parsi-
mony analyses contained 712 aligned sites of
which 230 were informative.
The basal node of the phylogenetic tree is com-
posed of a trichotomy (Fig. 14): (1) R. montagnei
from South Africa (100%); (2) R. decipiens and R.
peruensis from California (100%) and; (3) all of
the Macaronesian species in a well-supported
monophyletic group (90%). The Macaronesian
group is divided into two larger groups: Roccella
allorgei,R. fuciformis and R. maderensis (67%)
and; Roccella phycopsis through R. canariensis
(99%).
The species Roccella allorgei (100%), R. fuci-
formis (99%) and R. maderensis (97%) of the first
Macaronesian group are convincingly supported
as monophyletic entities (Fig. 14). The second
Macaronesian group is nearly fully supported
(99%) and contains R. phycopsis, R. elisabethae,
R. tuberculata, R. canariensis and R. tinctoria.
Within that group Roccella phycopsis makes up
the sister group to R. elisabethae, R. canariensis,
R. tinctoria and R. tuberculata (80%). At the next
level Roccella tuberculata constitutes the sister
group to the rest (61%) and finally R. elisabethae
is the sister group to R. canariensis and R. tincto-
ria (90%). The latter two taxa, regarded as a so-
called species pair, did not form distinct lineages
and were not supported as distinct species.
Discussion
By placing the molecular tree (Fig. 14) in a mor-
phological context some interesting observations
can be made. Several important morphological key
characters correspond very well with major nodes
of the jackknife tree obtained from the molecular
data. With the exception of the sterile and fertile
taxa of the species pair Roccella tinctoria/R. ca-
nariensis, all studied species including the newly
described Roccella elisabethae are clearly identi-
fied by the ITS data. In the future when the major-
ity of the Roccella species have been morphologi-
cally studied, and hypotheses of character homol-
ogy established, morphological characters will
ideally be included in the data set and analysed for
character congruence together with the molecular
data.
The Macaronesian species are well (90%) sup-
ported as monophyletic (Fig. 14), but there are no
obvious morphological features to distinguish
them from Roccella outside that region. Many
more Roccella species remain to be examined and
at this stage of the investigation it cannot be ex-
cluded that one or more species from outside Mac-
aronesia will nest within the Macaronesian group.
Within the Macaronesian group all species of
the fully supported clade Roccella phycopsis
through R. canariensis share a thallus morphology
with terete and more or less foveaete branches
(Figs 3, 4, 9, 11, 12). Thallus morphology thus
supports the idea that Roccella phycopsis, R. elis-
abethae, R. tuberculata, R. tinctoria and R. canar-
iensis are monophyletic and may have evolved in
the Macaronesian region. The four species Roc-
cella canariensis, R. elisabethae, R. tinctoria and
R. tuberculata are resolved in a monophyletic
group with 80% support (Fig. 14). Morphology re-
inforces the support of that grouping by the pres-
ence of some unique cortex features shared by the
species included. They differ from other Roccella
species by the presence of a cortex plectenchyma
where the hyphae at least in the upper part of the
cortex are somewhat interlaced and sometimes
squeezed together (Figs 5, 13). That type of cortex
has been termed the "canariensis-type" as opposed
to the "tinctoria-type" in which the cortex hyphae
are much more conspicuously arranged in pali-
sades (Follmann 1969). However, the terminology
is inappropriate and should be abandoned since
both Roccella canariensis and R. tinctoria have
the "canariensis-type" of cortex.
Roccella canariensis, R. tinctoria and R. tuber-
culata develop a unique, epicortical, non-textural,
surface gel overlaying the cortex plectenchyma
(Fig. 13). The gel conglutinates the hyphae and

424 Anders Tehler et al.
Symb. Bot. Ups. 34:1
Figure 14. Phylogeny of the Roccella species from Macaronesia and Europe. Parsimony jackknife frequencies a
t
supported nodes. Names in bold indicate fertile specimens.

The phylogeny and taxonomy of Roccella 425
Symb. Bot. Ups. 34:1
sclerotinizes the upper part of the cortex making
the whole thallus stiff. Furthermore, the gel is im-
penetrable to the hypochlorite solution so that the
thallus surface reacts C negative in a spot test. The
cortex below the epicortex otherwise reacts C+ red
but that becomes visible only if the epicortex gel is
physically removed. Roccella elisabethae lacks
the epicortical surface gel but is nevertheless
nested within the epicortical group (Fig. 14). Thus,
a loss of the epicortex character must be postulated
for R. elisabethae although the inclusive node is
only weakly supported by ITS data (61%).
Specimens of the species pair Roccella tincto-
ria/R. canariensis group are well supported (90%)
but the two taxa were not further resolved into sep-
arate clades. This particular species pair was also
investigated by Myllys et al. (1999b) who sug-
gested they were perhaps better treated as conspe-
cific since they phylogenetically did not divide
into homogenous groups with fertile and sterile
specimens. In the present analysis the Roccella
tinctoria/R. canariensis clade contains four sub-
groups in a polytomy with altogether six single
specimens. All subgroups contain a mixture of
sterile and fertile specimens, thus, corroborating
the conclusion of Myllys et al. (1999b).
From an evolutionary point of view it seems
reasonable to believe that sorediate individuals
arise from fertile individuals from time to time
(Tehler 1982) and that occasionally such sterile in-
dividuals may successfully continue to disperse
and establish clones that become distributed over
wide geographical areas. The results from the
present analysis of the Canarian populations of R.
canariensis (fertile) and R. tinctoria (sterile) indi-
cate such a situation, judging from the subgroups,
which contain sterile specimens (clones) that are
more closely related to fertile specimens than they
are to other sterile specimens or vice versa (Fig.
14). These facts notwithstanding, we would prefer
to do some complementary studies including data
from other molecular markers before we take fur-
ther steps to formally unite the two taxa. The main
reason for the precaution is that the two morphs of
R. canariensis/R. tinctoria are nearly obligately ei-
ther sterile or fertile. Both morphs are very com-
mon in Macaronesia and they are very well repre-
sented in herbaria collections. In spite of that, and
although AT have actively but without success
searched for intermediates in the field, we have
seen only one specimen of R. canariensis/R. tinc-
toria that unequivocally carried both soralia and
apothecia [Canary Islands, Hierro, 1987, Feuerer
HBG].
The putative species pair contained under Roc-
cella maderensis show the same pattern as the ac-
tual species pair R. canariensis/R. tinctoria, i.e.
sequences were not identical and there was an in-
ternal grouping with the fertile specimen nested in
the midst of the sterile ones. But even though inter-
mediate specimens carrying both soralia and
apothecia are very unusual they are occasionally
found. Furthermore, R. maderensis is a much more
rare species than R. canariensis/R. tinctoria and
inter- or intraspecific variation is not easily stud-
ied. Therefore, in this case we will not attempt to
distinguish two taxa before more evidence to the
contrary is at hand. The correct name for the com-
bined fertile/sterile taxon is Roccella maderensis.
The two morphs of Roccella allorgei also meet
all necessary criteria of being considered a so-
called species pair (Poelt 1970, Poelt 1972). One
morph is sorediate and with clones established
over wide geographical areas. The other morph is
obligately or near obligately fertile but not as com-
mon as the sorediate counterpart and its popula-
tions have more restricted distributions. The fertile
and sterile specimens of R. allorgei have identical
sequences lending support to the conclusion that
these two morphs are really conspecific and do not
merit any taxonomic status (see the species de-
scriptions above for more details). The correct
name for the combined fertile/sterile taxon is Roc-
cella allorgei. The fact that the sequences of both
the sorediate and fertile specimens of Roccella al-
lorgei are identical, form a contrast to the situation
described in the two other so-called species pairs
in this work, Roccella canariensis/R. tinctoria and
R. maderensis (fertile)/R. maderensis (sterile).
Consider the event that the fertile counterpart of
a species pair population become extinct and the
generative production of new sorediate individuals
discontinued. In such a case the already estab-
lished clones could be expected to form relational

426 Anders Tehler et al.
Symb. Bot. Ups. 34:1
patterns with each other as a result of being devel-
oped (budded off) from sexual parental popula-
tions at different times in the evolutionary history.
With that in mind one would expect to see some
variation and resolution between specimens of the
obligately sterile Roccella tuberculata and the
near obligately sterile R. phycopsis and R. fuci-
formis, but this is not the case despite the large
distances between the populations of the Azores,
Canary Islands and Estremadura. Sequences from
specimens within Roccella tuberculata show none
or negligible variation and the same situation is
true for R. phycopsis and R. fuciformis. An expla-
nation would be that the most competitive clone of
respective species has over time successfully taken
over and established in all of the Canary Islands,
Azores and Estremadura.
ITS sequences from the sterile Roccella tincto-
ria specimens showed some interesting features
well worth noting. The specimens from the Azo-
rean population showed very little or no sequence
variation but the sequences are markedly different
from those obtained from specimens from the Ca-
naries or Estremadura. The one exception is a
specimen from La Gomera, which is almost iden-
tical to those of the Azorean population. Hypothet-
ically, all Roccella tinctoria specimens found in
the Azores could be the result of a single dispersal
event from the abovementioned clone in the Ca-
nary Islands. In contrast to the homogenous Azo-
rean population, the specimens from both the Ca-
naries and Estremadura show a much higher de-
gree of within-population sequence variation. This
variation may be due to the fact that the fertile
Roccella canariensis is common in the Canarian
and the Estremadurean populations. Genetic vari-
ability between chronologically conserved clones
could be expected to be higher considering that the
sexually interbreeding parental populations con-
tinuously evolve and change their genetic constitu-
tion over time (cf Tehler 1982). These necessary
conditions are lacking from the Azores since the
fertile counterpart, Roccella canariensis, is absent
or very rare and thus genetic recombination will
only seldom take place.
The clade including the species Roccella allor-
gei,R. fuciformis and R. maderensis is not very
well supported although the species are habitually
somewhat similar, often having thalli with more or
less flattened branches. One of the questions we
asked concerning two species in this clade was, if
a feature such as the mere location of specific sec-
ondary substances in different parts of the thallus
could be used for distinguishing species, in this
case R. fuciformis and R. maderensis.Roccella fu-
ciformis and R. maderensis are morphologically
two confusingly similar species, in the field often
indistinguishable. Normally, the easiest way to dis-
tinguish them is by using hypochlorite solution (C)
to detect the chemically closely related orcinol
type paradepsides erythrin and lecanoric acid.
When applied to the thallus surface and soralia
both react C+ red in Roccella maderensis but in R.
fuciformis only the soralia reacts C+ red whereas
the thallus reacts C negative. However, the reaction
is occasionally faint or in parts of the thallus even
negative in some individuals of R. maderensis and
in R. fuciformis the thallus may sometimes react
faintly C+ red. Thus spot test identification of the
two species is not fully reliable. The K reaction is
usually useful to distinguish the two species. The
thallus surface of Roccella fuciformis normally re-
acts faintly K+ yellowish (lepraric acid) whereas in
R. maderensis it normally reacts K negative or
slightly K+ pinkish/reddish, when applied to sor-
alia or the outermost tips of the thallus branches
(montagnetol). As with the C reaction, the K reac-
tion is somewhat variable and rather often vague
and indistinct. The overlapping variability between
the two species and the fact that they are joined into
a node with 100% support, gave substance to the
hypothesis that they should actually be regarded as
conspecific and that the observed chemical varia-
tion was merely a case of infraspecific variation. As
evident from the present study, though, sequence
data from the ITS regions indicated that the two
taxa could very well be distinguished as species
since they were both clearly derived into separate
clades with very high support. Roccella fuciformis
received 99% support and R. maderensis received
97% support. The new HPLC investigation also
indicated that lepraric acid is characteristic to Roc-
cella fuciformis while montagnetol is characteris-
tic to R. maderensis. Both species contain erythrin

The phylogeny and taxonomy of Roccella 427
Symb. Bot. Ups. 34:1
but only Roccella maderensis produce lecanoric
acid, a fact that could be responsible for the more
commonly found C+ red thallus reaction in R. ma-
derensis. Conclusively, both the molecular data
and the new chemical data presented here suggest
that R. fuciformis and R. maderensis should be
treated as separate species, not as chemical strains
within species variation. The situation is further
complicated because Roccella maderensis may
also be very similar to R. allorgei, the most closely
related species to the sister pair R. maderensis and
R. fuciformis. They both share the C+ red thallus
reaction described above as well as the K+ pinkish/
reddish reaction. Usually Roccella maderensis has
more conspicuously flattened branches than R. al-
lorgei, which generally has more terete branches.
When morphologically converging R. maderensis
and R. allorgei may be confusingly similar and
very hard to distinguish. A more detailed study in-
cluding additional data and a larger sample of spec-
imens from the three species Roccella allorgei, R.
fuciformis and R. maderensis would certainly be
desirable.
Acknowledgements
We want to thank Dr Jens Klackenberg who has
revised the latin diagnosis. This study was finan-
cially supported by the Swedish Research Council
(VR grant 621-2002-4058) for Arthoniales phyl-
ogeny.
References
Awasthi, D. D. 1981. The typification of Roccella mon-
tagnei. – Bryologist 84: 216–219.
Darbishire, O. V. 1897. Über die Flechtentribus der Roc-
cellei. – Ber. Deutsch. Bot. Ges. 15: 2–10.
Darbishire, O. V. 1898a. Monographia Roccelleorum. –
Biblioth. Bot. 45: 1–102.
Darbishire, O. V. 1898b. Weiteres über die Flechtentri-
bus der Roccellei. – Ber. Deutsch. Bot. Ges. 16: 6–
16.
de Lamarck, J. B. & de Candolle, A. P. 1805. Flore
Française, ou descriptions succinctes de toutes les
plantes qui croissent naturellement en France, 3(2). –
H. Agasse, Paris.
Des Abbayes, H. 1947. Lichens des îles Açores récoltés
en 1937 par V. et P. Allorge. – Rev. Bryol. Lichénol.
16: 105–112.
Farris, J. S. 1997. Xac: Parsimony Jackknifer. – Mole-
kylärsystematiska laboratoriet, Naturhistoriska riks-
museet, Stockholm.
Feige, G. B., Lumbsch, H. T., Huneck, S. & Elix, J. A.
1993. The identification of lichen products by a
standardized high-performance liquid chromato-
graphic method. – J. Chromatogr. 646: 417–427.
Follmann, G. 1969. Roccella. – In: Poelt, J. (ed.). Bes-
timmungsschlüssel europäischer Flechten, Cramer,
Vaduz, pp. 704–706.
Follmann, G. 1995. On the impoverishment of the lichen
flora and the regression of the lichen vegetation in
coastal central and northern Chile during the last
decades. – Cryptog. Bot. 5: 224–231.
Follmann, G. 2001. An integrated key to, and a critical
survey of, the South American representatives of the
lichen family Roccellaceae (Arthoniales). – J. Hat-
tori Bot. Lab 90: 251–267.
Follmann, G. 2002. South America as diversity centre of
the lichen family Roccellaceae (Arthoniales). – Mitt.
Inst. Allg. Bot. Hamburg 30–32: 61–77.
Follmann, G. & Mies, B. 1988. Zur Kenntnis der Flech-
tenflora und Flechtenvegetation der Kapverdischen
Inseln. VI. Zusammenstellung von den Kapver-
dischen Inseln erstmalig beschriebener Flechtensip-
pen (Lichenophyta). – Courier Forschungsinst.
Senckenberg 105: 57–65.
Follmann, G. & Sánchez-Pinto, L. 1987. Zur Kenntnis
der Flechtenflora und Flechtenvegetation der Ka-
pverdischen Inseln III. Neue Flechtenfunde und ihre
Bedeutung für die Floren- und Vegetationsges-
chichte. – Courier Forschungsinst. Senckenberg 95:
175–188.
Gardes, M. & Bruns, T. D. 1993. ITS primers with en-
hanced specificity for basidiomycetes – application
to the identification of mycorrhizae and rusts. –
Molec. Ecol. 2: 113–118.
Grube, M. 1998. Classification and Phylogeny in the Ar-
thoniales (Lichenized Ascomycetes). – Bryologist
101: 377–391.
Howe, R. H. J. 1912. The lichens of the Linnean herbar-
ium with remarks on Acharian material. – Bull. Tor-
rey Bot. Club 39: 199–203.
Jørgensen, P. M., James, P. W. & Jarvis, C. E. 1994. Lin-
naean lichen names and their typification. – Bot. J.
Linn. Soc. 115: 261–405.
Montagne, C. 1857. Plantes cellulaires nouvelles, tant
indigènes qu’exotiques. – Ann. Sci. Nat. Bot. 7:
134–153.
Myllys, L., Källersjö, M. & Tehler, A. 1998. A compari-
son of SSU rDNA data and morphological data in
Arthoniales (Euascomycetes) phylogeny. – Bryolo-
gist 101: 70–85.
Myllys, L., Lohtander, K., Källersjö, M. & Tehler, A.
1999a. Applicability of ITS data in Roccellaceae
(Arthoniales, Euascomycetes) phylogeny. – Li-
chenologist 31: 461–476.
Myllys, L., Lohtander, K., Källersjö, M. & Tehler, A.

428 Anders Tehler et al.
Symb. Bot. Ups. 34:1
1999b. Sequence insertions and ITS data provide
congruent information on Roccella canariensis and
R. tuberculata (Arthoniales, Euascomycetes) phyl-
ogeny. – Molec. Phyl. Evol. 12: 295–309.
Poelt, J. 1970. Das Konzept der Artenpaare bei den
Flechten. – Vortr. Gesamtg. Bot. 4: 187–198.
Poelt, J. 1972. Die Taxonomische Behandlung von
Artenpaaren bei den Flechten. – Bot. Not. 125: 77–
81.
Santesson, R. 1944. Contributions to the lichen flora of
South America. – Ark. Bot. 31A: 1–28.
Staden, R. 1996. Staden sequence analysis package. –
Molec. Biotechnol. 5: 233–342.
Steiner, J. 1904. Flechten, auf Madeira und den Kanaren
gesammelt von J. Bornmüller in den Jahren 1900
und 1901. – Österr. Bot. Z. 9: 1–32.
Tehler, A. 1982. The species pair concept in lichenology.
– Taxon 31: 708–714.
Tehler, A. 1983. The genera Dirina and Roccellina (Roc-
cellaceae). – Opera Bot. 70: 1–86.
Tehler, A. 1990. A new approach to the phylogeny of
Euascomycetes with a cladistic outline of Arthoni-
ales focussing on Roccellaceae. – Canad. J. Bot. 68:
2458–2492.
Tehler, A. 2002a. Dirina. – In: Nash III, T. H., Ryan, B.,
Gries, C. & Bungartz, F. (eds), Lichen Flora of the
Greater Sonoran Desert Region. Volume I (the
pyrenolichens and most of the squamulose and mac-
rolichens), Lichens unlimited, Tempe, Arizona,
USA, pp. 178–180.
Tehler, A. 2002b. On the typification of two misunder-
stood Roccella names, R. tinctoria and R. fucoides. –
Taxon 51: 787–790.
Tehler, A. 2002c. Roccella. – In: Nash III, T. H., Ryan,
B., Gries, C. & Bungartz, F. (eds), Lichen Flora of
the Greater Sonoran Desert Region. Volume I (the
pyrenolichens and most of the squamulose and mac-
rolichens), Lichens unlimited, Tempe, Arizona,
USA, pp. 451–454.
Tehler, A. 2003. Roccella phycopsis the correct name for
R. fucoides (lichenized fungi). – Taxon 52: 651.
Tehler, A. & Källersjö, M. 2001. Parmeliopsis ambigua
and P. hyperopta (Parmeliaceae): species or chemo-
types? – Lichenologist 33: 403–408.
Tehler, A., Lohtander, K., Myllys, L. & Sundin, R. 1997.
On the identity of the genera Hubbsia and Reinkella
(Roccellaceae). – Symb. Bot. Ups. 32(1): 255–265.
Thompson, J. D., Gibson, T. J., Plewniak, F., Jean-
mougin, F. & Higgins, D. G. 1997. The ClustalX
window interface: flexible strategies for multiple se-
quence alignment aided by quality analysis tools. –
Nucl. Acids Res. 24: 4876–4882.
Vainio, E. A. 1901. Lichenes. 2. – In: Catalogue of the
African plants collected by F. Welwitsch in 1853–
1861. Cryptogamia, 2, British Museum Natural His-
tory, London, pp. 396–463.
White, T. J., Bruns, T. D., Lee, S. B. & Taylor, J. W.
1990. Amplification and direct sequencing of fungal
ribosomal RNA genes for phylogenetics. – In: Innis,
M. A., Gelfand, D. H., Sninsky, J. J. & White, T. J.
(eds), PCR Protocols. A guide to methods and appli-
cations, Academic Press, San Diego, pp. 315–322.