Intercontinental Mediterranean disjunct mosses: morphological and molecular patterns.

Abstract

This study focused on three species that occur disjunctly between western North America and the Mediterranean region of southern Europe, northern Africa, and western Asia, forming the so-called Madrean-Tethyan distribution pattern. Quantitative morphological characters were measured in New and Old World plants to find any subtle phenotypic differentiation between the disjunct populations. Sequences from the nuclear ribosomal internal transcribed spacer region were obtained from the same populations to assess differentiation at the molecular level and to compare molecular diversity with patterns of morphological similarity among plants. Little or no morphological differentiation existed between New and Old World plants in any of the species, but internal transcribed spacer (ITS) sequences revealed some phylogeographic structure. Patterns of morphological similarity in all three species were incongruent with phylogeographic structure revealed by sequence data. New World populations were more variable than Old World populations at the molecular level in the three species. Despite some evidence for differentiation between disjunct plants, no plausible mutation rate would date the divergence at ≥20 million years ago (MYA), as implied by the Madrean-Tethyan hypothesis. Recent long-distance dispersal is a more likely explanation for intercontinental disjunctions in these species.

Full text

Page 1

540
American Journal of Botany 90(4): 540–550. 2003.
INTERCONTINENTAL MEDITERRANEAN DISJUNCT
MOSSES: MORPHOLOGICAL AND MOLECULAR PATTERNS1
A. JONATHAN SHAW,2,4OLAF WERNER,3AND ROSA M. ROS3
2Department of Biology, Duke University, Durham, North Carolina 27708 USA; and
3Departamento de Biologı ´a Vegetal, Facultad de Biologı ´a, Universidad de Murcia, Campus de Espinardo, 30100-Murcia, Spain
This study focused on three species that occur disjunctly between western North America and the Mediterranean region of southern
Europe, northern Africa, and western Asia, forming the so-called Madrean-Tethyan distribution pattern. Quantitative morphological
characters were measured in New and Old World plants to find any subtle phenotypic differentiation between the disjunct populations.
Sequences from the nuclear ribosomal internal transcribed spacer region were obtained from the same populations to assess differ-
entiation at the molecular level and to compare molecular diversity with patterns of morphological similarity among plants. Little or
no morphological differentiation existed between New and Old World plants in any of the species, but internal transcribed spacer
(ITS) sequences revealed some phylogeographic structure. Patterns of morphological similarity in all three species were incongruent
with phylogeographic structure revealed by sequence data. New World populations were more variable than Old World populations at
the molecular level in the three species. Despite some evidence for differentiation between disjunct plants, no plausible mutation rate
would date the divergence at ?20 million years ago (MYA), as implied by the Madrean-Tethyan hypothesis. Recent long-distance
dispersal is a more likely explanation for intercontinental disjunctions in these species.
Key words:
Madrean-Tethyan; Scleropodium.
bryophyte phylogeography; Claopodium; cryptic speciation; Dicranoweisia; disjunctions; internal transcribed spacer;
Moss species tend to have much broader geographic ranges
than species of seed plants. Few mosses are cosmopolitan and
weedy, i.e., common in anthropogenic habitats and likely dis-
persed by human activity (e.g., Bryum argenteum Hedw., Fu-
naria hygrometrica Hedw.). Many species, in contrast, have
well-defined bi- or tricentric distributions that parallel disjunct
patterns long known in seed plants, for example, eastern North
America-eastern Asia (Iwatsuki, 1958), western North Amer-
ica-Europe (Schofield and Crum, 1972; Schofield, 1988). Such
seed plant disjunctions as these have been interpreted generally
in the context of previously continuous ranges disrupted by
climate change during the Tertiary (see, however, Wen, 1999).
Because many mosses exhibit similar patterns, it has been ar-
gued that disjunctions within moss species share a common
historical explanation with seed plant genera. Two corollaries
of this view are that many extant moss species are exceedingly
old and that they have been in a state of morphological stasis
for tens of millions of years (Crum, 1972).
This paper deals with mosses that are disjunct between the
Pacific coast of North America and the Mediterranean region
of southern Europe (and adjacent parts of Africa and western
Asia). Patterns of disjunction between southern Europe and
western North America have gained the attention of paleobot-
anists (Axelrod, 1973, 1975; Wolfe, 1975), systematists (Ra-
ven, 1973; Raven and Axelrod, 1978; Fritsch, 1996, 2001),
and ecologists (Cody and Mooney, 1978). Axelrod (1973,
1975) proposed a vicariance explanation for Mediterranean
seed plant disjunctions embodied in the Madrean-Tethyan hy-
pothesis. Accordingly, Mediterranean disjuncts reflect frag-
mentation of an ancestral, widespread, sclerophyllous flora that
1Manuscript received 26 July 2002; revision accepted 12 November 2002.
The authors thank Sandra Boles for technical assistance with all aspects of
the work. This research was supported by a collaborative grant to R. M. R.,
O. W., and A. J. S. from the Commission for Cultural, Educational, and Sci-
entific Exchange between the United States and Spain, and NSF DEB-
0075611 to A. J. S.
4Author for reprint requests (Fax: 919-660-7293, e-mail: shaw@duke.edu).
existed across western North America and Eurasia under sub-
humid conditions until the end of the Oligocene, about 25–20
million years ago (MYA). This theory has been challenged on
paleobotanical grounds because of limited evidence that scler-
ophyllous vegetation ever formed a continuous band between
presently disjunct regions (Wolfe, 1975). Inferences from mo-
lecular biogeographic analyses of extant taxa have shed new
light on this question. Divergence dates estimated from DNA
are consistent with Tertiary vicariance in Styrax (Fritsch,
2001), Arbutus (Hileman et al., 2001), and Datisca (Liston et
al., 1989a, 1992; Liston, 1997), although inferences from iso-
zymes are sometimes in conflict (Fritsch, 1996).
This paper focuses on three species of moss, Claopodium
whippleanum (Sull.) Ren. & Card., Dicranoweisia cirrata
(Hedw.) Lindb., and Scleropodium touretii (Brid.) L.F. Koch,
that are disjunct between the west coast of North America
from California to British Columbia or Alaska and the Med-
iterranean region of southern Europe, northern Africa, and
southwestern Asia. Species disjunct between Europe and the
Pacific coast of North America account for approximately 5%
of the mosses and 7% of the liverworts of Pacific North Amer-
ica (Schofield, 1984). While liverworts with this pattern tend
to be widespread in Europe, the mosses are more typically
Mediterranean (i.e., regions with hot dry summers and wet
winters). Although infraspecific disjunctions of this magnitude
are known in angiosperms (e.g., Styrax—Fritsch, 1996, 2001;
Senecio—Liston et al., 1989b), more often the disjunction is
reflected in the distribution of congeneric species or groups of
species (e.g., Arbutus—Hileman et al., 2001). The mosses con-
sidered in this paper, in contrast, appear to be morphologically
undifferentiated between continents, and the North American
and European populations have never been proposed to rep-
resent different species. If their disjunct patterns reflect ancient
vicariance dating to the Oligocene, morphological evolution
has remained at a virtual standstill for some 20 or more million
years.
The three species considered in this paper are fairly com-

Page 2

April 2003]541SHAW ET AL.—MEDITERRANEAN DISJUNCT MOSSES
mon in live oak or dry Douglas-fir forests in coastal areas of
California northward to southern Canada. In southern Europe
and northern Africa, they occur in ecologically similar zones
and, although not rare, may be less common than in western
North America. (Claopodium whippleanum is in fact rather
rare in southern Europe.) They do not occur in the eastern
United States, but isolated records exist for Dicranoweisia cir-
rata eastward in North America to Idaho and also in Mexico.
They are not, in any event, widespread species in the Northern
Hemisphere but rather have highly disjunctive distributions
that are essentially bicentric. The purposes of this study were
(1) to determine if subtle morphological differentiation exists
between North American and European populations and (2) to
assess cryptic differentiation between disjunct population sys-
tems using DNA sequence data from the nuclear ribosomal
internal transcribed spacer (ITS) region. Although divergence
times could not be reliably estimated in the absence of a fossil
record for calibration, we aimed to test the Madrean-Tethyan
prediction that intercontinentally disjunct populations are
clearly divergent at the molecular level.
MATERIAL AND METHODS
Study organisms—Claopodium whippleanum (Leskeaceae) occurs from
California to Alaska (Lawton, 1971) and grows on mesic to xeric soils at the
bases of trees, on humus or rotting wood, and on exposed soil banks. Ga-
metophytes are unisexual and sporophytes occur occasionally in western
North American populations. In the Mediterranean area, it is a rare species
known from some countries and islands of southwest Europe (Du ¨ll, 1985,
1992) and northern Africa (Jime ´nez et al., 2002). It grows on humid screes
in Quercus rotundifolia and Pinus forests with acidic soils. It is rarely found
with sporophytes. Eighteen populations were sampled for morphological and
molecular analyses (listed in the Appendix, http://ajbsupp.botany.org/v90/).
These include five populations from the Old World and 13 from North Amer-
ica.
Dicranoweisia cirrata (Dicranaceae) generally grows as small round col-
onies on rocks, in open live oak forests, or sclerophyllous chaparral in the
United States and in comparable microsites within coniferous forests north-
ward to Alaska. It is also known from Arizona and northern Mexico (Sharp
et al., 1994). In Europe D. cirrata is widespread northward to southern Scan-
dinavia. It is also known from the Macaronesian Islands, southwest Asia, and
northern Africa (Hill et al., 1992), where it usually forms cushions on tree
trunks, especially at the base of conifers. The species has bisexual game-
tophytes (autoicous; separate male and female buds), and plants commonly
produce sporophytes. Thirty-two populations were included in the analysis,
10 from Old World and 22 from North America (see Appendix at http://
ajbsupp.botany.org/v90/).
Scleropodium touretii (Brachytheciaceae) grows in the same sorts of hab-
itats as Claopodium whippleanum in western North America, but is more
common. It typically occurs on soil banks and sometimes rocks on a thin
layer of soil. It has unisexual gametophytes, and sporophytes are produced
occasionally. In Europe it is present in the Mediterranean, central and western
parts of the continent, north to Scotland and Denmark, also in Macronesia,
northern Africa, and southwestern Asia (Hill et al., 1994). Its ecology is very
similar to that in North America. Our sampling included six populations from
the Old World and seven from western North America.
Morphological measurements—Analyses of specimens we collected were
supplemented with specimens in DUKE and MUB. Additional North Amer-
ican collections were provided by J. Shevock (National Forest Service and
UC), W. B. Schofield (UBC), and F. Lara (Universidad Auto ´noma de Madrid).
Sampling for morphological and molecular analyses utilized different plants
from the same collection. Except where noted, vouchers for the analyses are
preserved in DUKE and MUB; in the case of loans from UC and UBC, the
original collections are preserved in those herbaria with a single stem used
for the analyses preserved at DUKE.
Claopodium whippleanum and Scleropodium touretii—Morphological mea-
surements included both continuous and discrete characters. Quantitative char-
acters measured for each species are listed in Table 1. For growth form, type
of branching, position of leaves on the stem, and a general impression of the
voucher was considered. Two microscopic slides were prepared with about
15–20 leaves each, one with leaves of the primary or secondary stem and the
other with branch leaves. For each quantitative leaf character, 10 measure-
ments were made from these preparations always taking length and width
from the same leaf. Leaf apex length was estimated by measuring the distance
from the base to the apex (defined by where the angle of the leaf margin
changed fairly abruptly, to the tip of the leaf). For leaf cell dimensions, 10
cells were measured, always taking length and width from the same cell.
Upper cells were measured in the upper third of leaves, basal cells in the one
or two most basal rows of cells in the leaf, and alar cells in the outer angles
of leaves. Because branch and stem leaves are different in size and shape in
Claopodium and Scleropodium (as in many other pleurocarpous mosses), leaf
and cell measurements were made on both types. Sporophyte features were
not taken into account for these species because of their scarcity in the col-
lections studied.
Dicranoweissia cirrata—For gametophyte size, the lengths of 10 plants
were measured, and one of the plants was selected for additional measure-
ments of leaf and cell dimensions. A microscopic slide was prepared with
about 15–20 leaves from the upper third of the stem. For leaf length and
width, and costa width, 10 leaves were measured. Cell dimensions were taken
from all the leaves in the slide preparation. When present, sporophytes were
described, but because of their scarcity a maximum of five were measured.
Morphological analyses—Analyses to detect geographic patterns in mor-
phological variation included both uni- and multivariate approaches. Means
from multiple measurements per population (previous section) were estimated
for use in the statistical analyses comparing Old and New World samples.
Within each species, populations were classified as Old or New World, and
significant differentiation in each morphological trait was tested separately
using univariate analyses of variance (ANOVAs). The ANOVA models spec-
ified region (Old vs. New World) as the main effect with error variances
derived from populations within these regions.
Multivariate patterns of variation were assessed using principal components
analyses (PCA) and UPGMA cluster analyses. For the PCA, plots of popu-
lations within each species in relation to the first two principal components
were constructed with populations designated as New or Old World (N and
o in Fig. 1). Only characters with no missing data for any of the populations
are shown in the plots. (For example, gemmae characters in Dicranoweisia
are not included in the PCA because some populations lacked them.) All
characters were included in UPGMA cluster analyses for each species (with
missing data coded as appropriate). Statistical analyses were conducted in
SAS (SAS Institute Inc., Cary, North Carolina, USA).
DNA extraction, PCR amplification, and sequencing—DNA extractions
followed a modification of Doyle and Doyle’s (1990) protocol, and nuclear
ribosomal DNA (nrDNA) was amplified as in Shaw (2000). Our nrDNA se-
quences include the last 5–25 bases of the 18s ribosomal RNA gene, ITS-1,
5.8s rRNA gene, ITS-2, and the first few bases of the 26s rRNA gene. Bound-
aries of the subregions were identified by comparison to available GenBank
sequences. Multiple ITS haplotypes were detected within plants from two
populations of Scleropodium (MED3, SP7), and two haplotypes were identi-
fied from each based on cloned sequences. Initial polymerase chain reaction
(PCR) products were cloned using the TOPO-TA cloning kit. A 0.8-?L sample
of fresh PCR product was mixed with 0.2 ?L of salt solution and 0.2 ?L of
vector and incubated at room temperature for 30 min. The ligation/cloning
reaction was added to 10–15 ?L of competent cells and incubated on ice for
30 min. The cells were heat shocked for 30 s at 42?C, transferred to ice, and
150 ?L SOC medium (at room temperature) was added to the cells. The

Page 3

542 [Vol. 90AMERICAN JOURNAL OF BOTANY
TABLE 1.Morphological characters in New and Old World populations of Claopodium whippleanum, Dicranowiesia cirrata, and Scleropodium
touretii (means ? 1 SE). *P ? 0.05; **P ? 0.01; ns ? not significant. All units in micrometers.
CharacterOld WorldNew World Significance level
Claopodium whippleanum
Stem leaf length
Stem leaf width
Stem leaf costa width
Stem leaf apical cell length
Stem leaf cell length
Stem leaf cell width
Branch leaf length
Branch leaf width
Branch leaf costa width
Branch leaf apical cell length
Branch leaf cell length
Branch leaf cell width
Branch leaf basal cell length
Branch leaf basal cell width
Dicranoweisia cirrata
Plant height
Leaf length
Leaf width
Costa width
Upper cell diameter
Basal cell length
Basal cell width
Seta length
Capsule length
Capsule diameter
Peristome length
Spore diameter
Gemma length
Gemma diameter
Scleropodium touretii
Stem leaf length
Stem leaf width
Stem leaf apical cell length
Stem leaf costa length (% leaf)
Stem leaf costa width
Stem leaf upper cell length
Stem leaf upper cell width
Stem leaf basal cell length
Stem leaf basal cell width
Stem leaf alar cell length
Stem leaf alar cell width
Branch leaf length
Branch leaf width
Branch leaf apical cell length
Branch leaf costa length (% leaf)
Branch leaf costa width
Branch leaf upper cell length
Branch leaf upper cell width
Branch leaf basal cell length
Branch leaf basal cell width
Branch leaf alar cell length
Branch leaf alar cell width
654.6 ? 54.3
309.5 ? 27.4
28.0 ? 2.1
20.0 ? 0.6
13.6 ? 0.6
5.6 ? 0.3
581.9 ? 44.0
236.0 ? 26.6
25.1 ? 1.8
20.0 ? 0.7
12.7 ? 0.4
5.9 ? 0.1
16.5 ? 0.6
6.3 ? 0.2
661.4 ? 29.5
298.0 ? 16.5
30.2 ? 1.3
21.8 ? 0.9
14.5 ? 0.6
5.5 ? 0.1
625.5 ? 29.6
243.3 ? 11.0
28.5 ? 1.3
20.0 ? 0.6
12.3 ? 0.3
5.7 ? 0.2
15.4 ? 0.7
6.5 ? 0.2
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
8.9 ? 1.1
1934.9 ? 71.9
373.6 ? 24.9
49.1 ? 1.1
9.3 ? 0.4
31.3 ? 0.9
15.6 ? 0.5
5.6 ? 0.6
1387.1 ? 80.0
588.8 ? 32.4
112.5 ? 3.5
17.9 ? 0.6
117.1 ? 11.7
40.5 ? 1.2
12.1 ? 0.6
2134.2 ? 56.6
379.5 ? 9.2
48.6 ? 1.4
10.0 ? 0.2
37.0 ? 1.6
14.1 ? 0.3
5.1 ? 0.3
1486.1 ? 54.4
559.1 ? 17.1
120.5 ? 4.4
17.3 ? 0.3
127.6 ? 14.2
46.9 ? 1.9
**
*
ns
ns
ns
*
*
ns
ns
ns
ns
ns
ns
*
1306.2 ? 59.8
760.9 ? 70.8
78.1 ? 17.6
65.4 ? 2.1
61.7 ? 8.6
57.5 ? 5.1
4.6 ? 0.2
28.4 ? 2.0
10.9 ? 0.4
22.9 ? 2.5
13.3 ? 0.7
1212.8 ? 54.8
655.2 ? 31.1
97.8 ? 21.4
72.4 ? 3.1
53.6 ? 3.5
54.5 ? 1.9
4.7 ? 0.3
27.0 ? 2.1
10.5 ? 0.4
24.6 ? 2.4
11.3 ? 0.7
1356.3 ? 86.0
767.9 ? 52.9
99.6 ? 33.5
60.0 ? 3.7
53.9 ? 6.6
60.2 ? 4.1
5.0 ? 0.3
29.3 ? 2.4
11.0 ? 0.6
23.4 ? 2.6
13.1 ? 0.7
1269.6 ? 88.6
665.4 ? 72.9
53.8 ? 13.5
73.8 ? 2.5
50.9 ? 6.4
54.8 ? 4.3
4.8 ? 0.3
28.1 ? 0.4
10.6 ? 0.4
20.1 ? 1.9
11.3 ? 0.6
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
mixture was shaken horizontally at 200 rpm for 1 h. The broth culture was
plated on agar (IPTG 0.5 mmol/L, X-gal 80 ?g/mL, tetracycline 12.5 ?g/mL,
ampicillin 100 ?g/mL) and incubated overnight at 37?C. Colonies containing
the vector with a PCR insert were used as template for a second PCR am-
plification.
Phylogenetic analyses—Sequence chromatograms were compiled using Se-
quencher software (version 2.0, Gene Codes, Ann Arbor, Michigan, USA) to
produce contigs based on nucleotide identifications from both DNA strands.
All sequences were aligned by eye, with gaps inserted where needed to pre-
serve nucleotide homology. Separate alignments were made for each species.
A total of five indels and one duplication were scored (0,1) across the three
species as follows; Claopodium, 2; Dicranoweisia, 1; Scleropodium, 3. Clao-
podium had a 17-base pair (bp) insertion/deletion (indel) in the ITS-1 and a
three-nucleotide (apparent) duplication (GTT) in ITS-2. The remaining indels
in other species were single-nucleotide differences in ITS-1.
The ITS region was sequenced for outgroup taxa in order to root phylo-
genetic trees from each study species: Claopodium crispifolium (Hook.) Ren
& Card. for C. whippleanum, Dicranoweisia crispula (Hedw.) Lindb. ex Milde
for D. cirrata, and Scleropodium obtusifolium (Jaeg.) Kindb. in Macoun &

Page 4

April 2003] 543SHAW ET AL.—MEDITERRANEAN DISJUNCT MOSSES
Kindb. for S. touretii. For Claopodium and Dicranoweisia, however, outgroup
sequences, although readily alignable, were sufficiently divergent relative to
ingroup sequences that the root could attach almost anywhere, yielding com-
pletely unresolved topologies in the strict consensus trees for rooted phylog-
enies. Consequently, a rooted tree is presented only for Scleropodium. Al-
though historical inferences for Claopodium and Dicranoweisia are limited
by having only unrooted networks, patterns of genetic diversity and differ-
entiation could still be investigated.
Because of the small sizes of the three data sets, exhaustive searches for
the most parsimonious trees were conducted in PAUP (Swofford, 2001). Trees
were produced in McClade (Maddison and Maddison, 1992) to reflect rela-
tionships implied by unweighted pair group method with arithmetic means
(UPGMA) clustering of morphological similarities in each species. These
trees were enforced as topological constraints in parsimony analyses of ITS
data. Most parsimonious (MP) trees under the constraints were compared to
unconstrained trees using the parametric Kishino-Hasegawa and nonparamet-
ric Templeton tests as implemented in PAUP. Additional topological con-
straints in which monophylly of North American vs. European populations
was forced were also compared to unconstrained ITS trees. The MP trees for
each species produced under such constraints were compared to the uncon-
strained trees.
Population genetic analyses—Estimates of nucleotide diversity and differ-
entiation between North American and European populations of each species
were conducted using SITES (Hey and Wakeley, 1997) and ARLEQUIN ver-
sion 2.00 (Schneider et al., 2000). Two estimates of theta (?) were calculated
separately for disjunct populations of each species to evaluate levels of mo-
lecular diversity in North American and Old World plants. The numbers of
pairwise differences between pairs of haplotypes yields an estimate of ??
(Tajima, 1983), and ?s is based on the numbers of segregating sites (Watter-
son, 1975). Patterns of differentiation were assessed by the number of fixed
differences between plants disjunct on different continents, the number of
shared polymorphic nucleotide sites, and genetic diversity among populations
(FST). The FSTand tests of its significance were computed according to Ex-
coffier et al. (1992), using ARLEQUIN (Schneider et al., 2000).
RESULTS
Morphological analyses—All specimens of Claopodium
whippleanum examined from both Old and New World pop-
ulations were very similar morphologically, with the exception
of population BC18 from Canada (Vancouver Island). Plants
in this population were erect rather than prostrate, as is typical
of this species, and the leaves were not conduplicate as in the
rest of specimens studied.
American specimens of Scleropodium touretii appeared to
show greater qualitative variability than those from Europe,
where they seemed homogeneous. However, a difference in
level of variation was not detected in statistical analyses of the
quantitative data.
The morphology of Dicranoweisia cirrata has not been ac-
curately described in floristic treatments. According to descrip-
tions in moss floras for Britain and Ireland (Smith, 1978),
Fenncoscandia (Nyholm, 1986), Mexico (Sharp et al., 1994),
and the Pacific Northwest of North America (Lawton, 1971),
D. cirrata has smooth upper laminal cells, 9–14 ?m wide, alar
cells not forming auricles or only slightly differentiated, and
lamina unistratose or bistratose at apex and margins. We ob-
served that this species is more variable than has been de-
scribed. For instance the upper cells range from (6–)8–13 ?m
wide, smooth in the Old World specimens but sometimes pa-
pillose in the American samples. The alar cells are sometimes
clearly differentiated in the American specimens but never in
the Old World. In both New and Old World plants, the upper
part of the lamina of the leaf can be uniformly bistratose, bi-
stratose in patches, or unistratose. The margins are consistently
bistratose or sometimes tristratose at least in the upper and
median part of leaf; they are only unistratose in the basal part
of leaf. Some of these characters have been described in the
related species, D. crispula (Hedw.) Milde, but not in D. cir-
rata.
There was no evidence of strong morphological differenti-
ation in quantitative traits between Old World and North
American populations of any species. For Claopodium and
Scleropodium, no character even approached significance,
whereas limited differentiation was suggested for Dicranow-
eisia (Table 1). North American plants of Dicranoweisia tend-
ed to be taller, with larger leaves, larger basal leaf cells, and
larger asexual gemmae. Despite these subtle differences in ga-
metophyte sizes between North American and European plants
of Dicranoweisia, comparable differences in sporophyte di-
mensions were not detected (capsule and peristome traits in
Table 1).
Additional subtle patterns of morphological variation were
noted in the gametophytes of Dicranoweisia. Five samples
from North America (BC37, OR40, CA103, CA1276, and
CA1277) are here referred to as NA2 to distinguish them from
the rest of the North American plants (NA1) and from the Old
World plants (OW). Plants from the NA2 populations were
taller (OW ? 8.95 mm, NA1 ? 11.75 mm, and NA2 ? 13.18
mm) with longer (OW ? 31.3 ?m, NA1 ? 33.7 ?m, and
NA2 ? 48.4 ?m) but narrower (OW ? 15.6 ?m, NA1 ? 14.6
?m, and NA2 ? 12.4 ?m) basal cells. The last two measure-
ments are related to the presence of auricles in these five sam-
ples; such auricles are absent in all Old World samples and
are only slightly developed in one sample from NA1. Papillae
are absent from the leaf cells in all samples from the Old
World, present in seven of 17 populations from NA1, and pre-
sent in all five populations of NA2. On the other hand, gem-
mae were present in six of 10 populations from the Old World,
in seven of 17 populations from NA1, and none of the five
populations from NA2. These morphological variations do not
correspond to any geographic pattern in North America, nor
are they reflected in the ITS data. (In fact, plants belonging to
the NA2 group of populations can be found on the widely
divergent branches of the ITS tree.) These five populations do
not belong to Dicranoweisia crispula, a related species that
has auriculate leaves and papillose leaf cells, but is differen-
tiated from D. cirrata by such other characters as plane leaf
margin, smaller midleaf cells, fine longitudinal cuticular stri-
ations on the leaf cells, and ovate to ellipsoid capsules.
An absence of general morphological differentiation be-
tween Old World and North American populations of the three
species was evident in the multivariate analyses. Neither the
PCA nor UPGMA cluster analyses yielded any hint of gen-
eralized differentiation (Figs. 1–4). In many instances, Euro-
pean and North American populations were most similar to
one another, and the phenograms (Figs. 2–4) suggest a random
pattern with regard to New and Old World populations. Within
these broader geographic regions, proximate populations did
not tend to be more similar than those separated by greater
distances. It is clear that morphological variation is not struc-
tured geographically to any significant extent in any of the
three species we investigated.
Molecular analyses—The 5.8 rRNA gene was 157 bp in
all three species. In Claopodium, ITS-1 was 236–254 bp with
29 variable nucleotide sites and one indel, and ITS-2 was 277–

Page 5

544 [Vol. 90AMERICAN JOURNAL OF BOTANY
Fig. 1.
first two principal components in a multivariate analysis of morphological characters. See Table 1 for lists of characters measured in each species.
Plots of New (N) and Old (o) World populations of Claopodium whippleanum, Dicranoweisia cirrata, and Scleropodium touretii in relation to the
279 bp with 30 variable sites and two indels. The 5.8s rRNA
gene was invariant. Two equally most parsimonious trees were
found in an exhaustive search of the ITS data (Fig. 2). The
five European populations were monomorphic in terms of ITS
sequences and were differentiated from North American pop-
ulations by a single mutation (Fig. 2). North American pop-
ulations, in contrast, were more variable, with 37 mutations
differentiating populations. Among the North American plants,
three nucleotide sites (two in ITS-1 and one in ITS-2) were
polymorphic for more than two nucleotides. Both ?? and ?s
were high for North American populations of Claopodium rel-
ative to the two other species (Table 2). Theta (?) was of
course zero for the monomorphic European populations. Only
about 20% of the nucleotide variation in Claopodium was at-
tributable to differentiation between Old and New World
plants, but FSTwas significantly greater than zero.
Two populations from California (CA1265, CA1268) shared
an ITS haplotype with a population from British Columbia
(BC17), and two other California populations (CA1266,
CA105) likewise shared a haplotype. Otherwise, all of the
American populations were unique, although no fine-scale
geographic patterns were evident among the populations.
When the phylogenetic analysis of ITS sequences was con-
strained to the branching topology suggested by UPGMA clus-
tering of morphological variation, tree length increased from
39 steps (unconstrained) to 65 steps, an increase of 67%, and
homoplasy indices increased substantially (Table 3). The con-
strained and unconstrained trees were significantly different
using both the Kishino-Hasegawa (parametric) and Templeton
(nonparametric) tests. Constraining the analyses to enforce
monophylly of North American vs. European populations did
not increase tree length.
In Dicranoweisia, ITS-1 was 291–292 bp with nine variable
nucleotide sites and two indels, and ITS-2 was 288 bp with
seven variable sites. One site was polymorphic in the 5.8s
rRNA gene. An exhaustive search of the ITS data yielded a
single most parsimonious tree with no homoplasy (Fig. 3, Ta-
ble 3). A group of 17 populations, including plants from both
North America and Europe, were monomorphic and identical
in terms of ITS sequences. This group is distinguished by at
least three mutations from five other populations, all Old
World. This latter group of Old World plants is distinguished
from all remaining populations, all American, by a single mu-
tation. Within the latter group of American plants, nine pop-

Page 6

April 2003] 545SHAW ET AL.—MEDITERRANEAN DISJUNCT MOSSES
Fig. 2.
a UPGMA phenogram based on morphological characters listed in Table 1.
Claopodium whippleanum. One of two most parsimonious networks (top, left) and the consensus network (top, right) from ITS sequence data and
ulations from the United States are differentiated by five mu-
tations from the Mexican population, which was most similar
to the European plants (Fig. 3). Constraining North American
and European populations as monophyletic groups increased
branch length by four steps over the unconstrained analyses.
The two topologies were significantly incongruent at P ? 0.05
(Kishino-Hasegawa and Templeton tests).
Constraining the phylogenetic analysis of the ITS data to
the branching topology revealed in the UPGMA phenogram
from morphological data increased branch length from 16
steps to 103 steps (500%) and, not surprisingly, estimates of
homoplasy increased equivalently (Table 3). Neither molecular
nor morphological data resolved European and North Ameri-
can populations as distinct groups, and grouping of popula-
tions was highly incongruent between data sets. California
populations CA1271 and CA1274, for example, were different
in terms of ITS sequences but were very similar morpholog-
ically (Fig. 3). In contrast, the Spanish populations SP30 and

Page 7

546[Vol. 90AMERICAN JOURNAL OF BOTANY
Fig. 3.
Dicranoweisia cirrata. Single most parsimonious network (top) and a UPGMA phenogram based on morphological characters listed in Table 1.
SP32 had identical ITS sequences but were relatively different
morphologically.
Within the group of North American plants distinguished by
five mutations, two subgroups differed by a single mutation,
and one population (CA1269) was characterized by two au-
tapomorphies. Otherwise, populations were monomorphic.
Taking into account all of the North American plants, however,
?? was more than twice as high in North American than Eu-
ropean plants, and ?s was also higher (Table 2). However,
standard deviations of these estimates were high and overlap-
ping. The FSTfor North American vs. European plants was
lower than in either of the other two species, but was signif-
icantly greater than zero.
In Scleropodium, ITS-1 was 286–289 bp with eight variable
nucleotide sites and four indels, and ITS-2 was 292 bp with
seven of them variable. Phylogenetic analyses of ITS sequenc-
es from Scleropodium yielded one most parsimonious tree
(Fig. 4). Two samples from the Old World, populations MO3

Page 8

April 2003] 547SHAW ET AL.—MEDITERRANEAN DISJUNCT MOSSES
Fig. 4.
morphological characters listed in Table 1.
Scleropodium touretii. Single most parsimonious tree, rooted with a sequence from S. obtusifolium (shaded) and a UPGMA phenogram based on
TABLE 2.Estimates of nucleotide diversity and differentiation within and between North American and European populations of disjunct species.
Population
???s
FST
Fixed
differences
Shared
polymorphic sites
Claopodium
Europe
North America
Europe ? North America
Dicranoweisia
Europe
North America
Europe ? North America
Scleropodium
Europe
North America
Europe ? North America
0.000
7.269 ? 4.093
0.000
8.3784 ? 3.484
0.1987 (P ? 0.007)10
2.422 ? 1.620
5.169 ? 2.903
2.121 ? 1.164
3.566 ? 2.903
0.1665 (P ? 0.04)03
1.036 ? 0.874
1.714 ? 1.296
0.385 ? 0.385
2.041 ? 1.223
0.6533 (P ? 0.0001)00
and SP7, yielded sequences that suggested intragenomic poly-
morphism. Subsequently, two ITS sequences differing in a sin-
gle nucleotide substitution were cloned from each of these
plants. One of the two cloned sequences from each plant
(clone 1 in Fig. 4) was identical to ITS sequences obtained
from other European plants of Scleropodium, whereas the oth-
er (clone 2) was uniquely shared by these two populations.
Clone 2 differed from the outgroup by only two mutations and
may represent an early diverging ITS sequence within S. tour-
etii (Fig. 4).
Inclusion of an outgroup sequence from a North American
population of S. obtusifolium in the phylogenetic analysis in-
dicates that the North American populations are nested within
a paraphyletic basal grade of European populations of S. tour-
etii. Clone 2 of the two polymorphic plants is most similar to
the outgroup sequence, and the widespread European ITS hap-
lotype differs by only one additional substitution. A population
from Tenerife (TEN1) shares a synapomorphic mutation with
the seven North American populations. The remaining North
American populations, with the exception of CA1259, are
united by 2 mutations (Fig. 4). The FSTbetween North Amer-
ican and European populations was higher than for Claopo-
dium and Dicranoweisia and was significantly greater than
zero. However, constraining the ITS analysis to enforce mon-
ophylly of the North American and European populations in-
creased tree length by only one step, and the two topologies
were not significantly incongruent. Constraining the ITS phy-
logeny to the topology obtained from the UPGMA analysis of
morphological data increased branch length from eight to 23
steps (190%), with an expected increase in homoplasy (Table
2). As in the previous two species, relationships implied by
morphological vs. molecular data were significantly incongru-
ent in terms of molecular evolution, as assessed by the Kish-
ino-Hasegawa test. Neither American nor Old World popula-

Page 9

548[Vol. 90AMERICAN JOURNAL OF BOTANY
TABLE 3.
strained analyses of ITS variation in Claopodium whippleanum,
Dicranowiesia cirrata, and Scleropodium touretii. HK and T ?
significance of Kishino-Hasegawa and Templeton tests, respective-
ly, of differences in tree length. CI ? consistency index; RI ?
retention index; RC ? rescaled consistency index. *P ? 0.05; **P
? 0.01; ***P ? 0.001; ns ? not significant.
Tree length and tree statistics for constrained and uncon-
StatisticUnconstrained ConstrainedHK/T
Claopodium whippleanum
Length
CI
RI
RC
Dicranoweisia cirrata
Length
CI
RI
RC
Scleropodium touretii
Length
CI
RI
RC
39
0.769
0.750
0.577
65
0.462
0.028
0.013
**/***
16
1.000
1.000
1.000
103
0.155
0.094
0.015
**/***
8
1.000
1.000
1.000
23
0.348
0.250
0.087
*/ns
tions of Scleropodium are variable in the ITS region, and this
is reflected in low estimates for ? from nucleotide substitutions
alone (Table 3).
DISCUSSION
Despite observations of some subtle patterns of morpholog-
ical variation in qualitative characters that appear to have a
geographic component, quantitative measurements of morpho-
logical characters did not reveal significant differentiation be-
tween Old and New World populations in any of the three
species we studied. Some limited amount of differentiation,
mainly in gametophyte size, was revealed within Dicranow-
eisia, but the differences accounted for a minimal proportion
of the overall variation and were barely significant.
Molecular data, in contrast, did reveal some geographic pat-
terns in both cladistic structure and levels of molecular diver-
sity. Relationships among populations implied by morpholog-
ical patterns are significantly incongruent with genealogical
relationships inferred from molecular data. For Claopodium
and Dicranoweisia, ITS phylogenies were unrooted so it was
not possible to assess monophyly of infraspecific clades. Nev-
ertheless, the five Old World populations of Claopodium
shared a mutation that set them apart from all New World
populations. It is possible that rooting the network would re-
solve both Old and New World populations as monophyletic,
or, alternatively, populations of one of the two regions may
represent a basal grade that gave rise to a disjunct monophy-
letic group of populations.
In Dicranoweisia, neither the Old nor New World popula-
tion forms a monophyletic group. Plants belonging to a clade
of North American plants share five mutations, but other North
American populations are identical to Old World plants.
Again, in the absence of a rooted phylogeny, it is difficult to
interpret this pattern in a historical context. Intercontinental
dispersal might be the basis for the disjunct distribution of ITS
haplotypes, or shared haplotypes between New and Old World
plants may have resulted from retention of ancestral polymor-
phism. The well-marked clade of North American plants could
be derived from plants with such ancestral haplotypes. This
interpretation would imply that the group of 17 New and Old
World populations that share a common ITS haplotype are
basal and that other Old World plants (i.e., populations
TUR31, GRE 91, MO90, TEN33, and MOR88) are derived
from them. The haplotype network, polarized in this way, fur-
ther implies that the species spread to Mexico, and then north-
ward into the United States and Canada.
The molecular pattern in Scleropodium is more like that
observed in Claopodium. New and Old World populations are
differentiated, but one population from California lacks the
mutations that otherwise characterize New World plants, and
a population from Tenerife shares that same mutation that is
otherwise limited to New World plants. The rooted phylogeny
for Scleropodium touretii suggests that the Old World plants
form a paraphyletic basal grade from which the New World
plants were derived. Indeed, plants from two Old World pop-
ulations (one from Spain, one from Morocco) had two differ-
ent ITS haplotypes, and one of them differed from that found
in the outgroup (S. obtusifolium) by only two mutations. Most
of the infraspecific ITS sequence variation in Scleropodium
was attributable to differentiation between North American
and European plants (i.e., FST? 0.6533), although the overall
level of variation was lower than in the other two species.
It may be significant that in all three species, North Amer-
ican populations contained higher levels of nucleotide diver-
sity than did their Old World conspecific counterparts. In Clao-
podium, Old World populations were monomorphic and fixed
for a single ITS haplotype, whereas New World populations
were characterized by numerous nucleotide substitutions. Sim-
ilarly, in both Dicranoweisia and Scleropodium, New World
populations were characterized by higher levels of ?, although
high standard deviations associated with estimates made these
differences not statistically significant. The highest ? value
was from North American populations of Claopodium, fol-
lowed by American populations of Dicranoweisia, then Scler-
opodium. The values of ? from Old World populations were
uniformly low or, as in Claopodium, zero.
Nucleotide diversity estimated by ? can reflect a number of
biological attributes. Population size (at present as well as in
the recent past), sex ratios, generation time, and age of the
species are among those factors that could underlie differences
in standing levels of molecular diversity. There are no obvious
differences in population structure (large vs. small breeding
populations) or generation time between Old and New World
plants in any of the three mosses. Sex ratios were not inves-
tigated. It could be argued that higher levels of molecular di-
versity in North American plants might result from longer oc-
cupation of the American parts of their ranges, but the rooted
phylogeny of Scleropodium touretii contradicts that interpre-
tation. It should be noted, however, that broader sampling
within Scleropodium might change interpretations of evolu-
tionary polarity within S. touretii. The pattern of differing lev-
els of nucleotide diversity is most obvious in Claopodium.
This species is more common in western North America than
in southern Europe, where it is quite rare. This pattern of abun-
dance, as well as the difference in molecular diversity, is con-
sistent with a more recent colonization of Europe from North
America, but in the absence of a rooted phylogeny, this hy-
pothesis is speculative.
Cryptic speciation, the evolution of phylogenetically distinct
taxa that are not marked morphologically, has been docu-
mented in both mosses and liverworts (Shaw, 2001). Mosses

Page 10

April 2003] 549SHAW ET AL.—MEDITERRANEAN DISJUNCT MOSSES
and liverworts are morphologically simple plants, and it is
perhaps not surprising that evolutionary differentiation is not
always evident from morphological characters. In this study,
subtle differentiation revealed by ITS sequences are not re-
flected in morphological variation among populations. Nev-
ertheless, none of the three species investigated here appear to
consist of cryptic sister species disjunct between New and Old
Worlds. We observed morphological variation within Dicra-
noweisia that is at odds with published descriptions of this
species in regional floras, and some of the features we ob-
served are more characteristic of the related species, D. cris-
pula. This latter species is distinguished by other characters
and our observations cannot be explained by misidentifica-
tions, nor by the occurrence of cryptic speciation within the
morphologically defined D. cirrata. Morphological variation
did not correspond to molecular differentiation in any way that
suggests cladogenesis.
Our results are not consistent with a Madrean-Tethyan ex-
planation for the disjunct distributions of Claopodium whip-
pleanum, Dicranoweisia cirrata, or Scleropodium touretii. We
were not able to assess absolute ages by calibrating rates of
molecular evolution in these species, but no plausible rate of
nucleotide substitution could yield a divergence time ?20–25
MYA. Even in Claopodium, where North American and Old
World plants are distinguishable by their ITS sequences, and
in Scleropodium, where, with two exceptions, a similar pattern
of differentiation was found, the limited degree of differenti-
ation almost certainly reflects more recent divergence. In Di-
cranoweisia, Old and New World plants neither form distinct
clades, nor differ to an extent that makes Tertiary vicariance
plausible. More recent long-distance dispersal is the most like-
ly explanation for the intercontinental disjunctions character-
izing these species.
There has been much discussion about infraspecific inter-
continental disjunctions in bryophytes. Because bryophyte
species exhibit many of the same patterns found in seed plants
at the generic level (i.e., vicariant species rather than popula-
tions) and because many such patterns are thought to have
historical bases dating as far back as the early Tertiary, these
patterns have been used to argue that many bryophyte species
are extremely old (e.g., Crum, 1972). A corollary of this in-
terpretation is that such species have remained morphologi-
cally unchanged for tens of millions of years. The western
North American–southern Europe disjunction as exemplified
by the three species included in this study is one such pattern
of disjunction in mosses. Our results, however, do not support
the interpretation that this disjunction is ancient and therefore
obviate any need to hypothesize morphological stasis over vast
amounts of time.
Bryophytes represent an underutilized group for biogeo-
graphic studies. In addition to widespread species with more
or less continuous distributions across the three northern con-
tinents, many patterns of infraspecific disjunction are ripe for
analyses based on newly available molecular methods. Recent
work on mosses disjunct between North America and Europe
(Shaw and Allen, 2000) demonstrate the efficacy of molecular
tools for studying bryophyte distributions. A substantial num-
ber of species are disjunct between eastern Asia and eastern
North America, but none has been subjected to molecular phy-
logeographic analyses. Our results suggest that when the utility
of morphological variation for biogeographic inferences is lim-
ited by homogeneity across continents, molecular data may
reveal patterns that permit additional insights into population
level processes such as dispersal and mating patterns.
LITERATURE CITED
AXELROD, D. I. 1973. History of the Mediterranean ecosystem in California.
In F. Di Cristi and H. A. Mooney [eds.], Mediterranean type ecosystems:
origin and structure, 225–277. Springer-Verlag, New York, New York,
USA.
AXELROD, D. I. 1975. Evolution and biogeography of the Madrean-Tethyan
sclerophyll vegetation. Annales of the Missouri Botanical Garden 62:
280–334.
CODY, M. L., AND H. A. MOONEY. 1978. Convergence versus nonconverg-
ence in Mediterranean-climate regions. Annual Review of Ecology and
Systematics 9: 265–321.
CRUM, H. A. 1972. The geographic origins of the mosses of North America’s
eastern deciduous forest. Journal of the Hattori Botanical Laboratory
35: 269–298.
DOYLE, J. J., AND J. L. DOYLE. 1990. Isolation of plant DNA from fresh
tissue. Focus 12: 13–15.
DU¨LL, R. 1985. Distribution of the European and Macaronesian mosses (Bry-
ophytina). Part II. Bryologische Beitra ¨ge 5: 110–232.
DU¨LL, R. 1992. Distribution of the European and Macaronesian mosses (Bry-
ophytina). Annotations and progress. Bryologische Beitra ¨ge 8/9: 1–223.
EXCOFFIER, L., P. E. SMOUSE, AND J. M. QUATTRO. 1992. Analysis of mo-
lecular variance inferred from metric distances among DNA haplotypes:
application to human mitochondrial DNA restriction data. Genetics 131:
479–489.
FRITSCH, P. 1996. Isozyme analysis of intercontinental disjuncts within Styrax
(Styracaceae): implications for the Madrean-Tethyan hypothesis. Ameri-
can Journal of Botany 83: 342–355.
FRITSCH, P. 2001. Phylogeny and biogeography of the flowering plant genus
Styrax (Styracaceae) based on chloroplast DNA restriction sites and DNA
sequences from the internal transcribed spacer region. Molecular Phy-
logenetics and Evolution 19: 387–408.
HEY, J., AND J. WAKELEY. 1997. A coalescent estimator of the population
recombination rate. Genetics 145: 833–846.
HILEMAN, L. C., M. C. VASEY, AND V. T. PARKER. 2001. Phylogeny and
biogeography of the Arbutoideae (Ericaceae): implications for the Mad-
rean-Tethyan hypothesis. Systematic Botany 26: 131–143.
HILL, M. O., C. D. PRESTON, AND A. J. E. SMITH. 1992. Atlas of the bryo-
phytes of Britain and Ireland, vol. 2, Mosses (except Diplolepidae). Har-
ley Books, Colchester, UK.
HILL, M. O., C. D. PRESTON, AND A. J. E. SMITH. 1994. Atlas of the bryo-
phytes of Britain and Ireland, vol. 3, Mosses (Diplolepidae). Harley
Books, Colchester, United Kingdom.
IWATSUKI, Z. 1958. Correlations between the moss floras of Japan and of the
Southern Appalachians. Journal of the Hattori Botanical Laboratory 20:
304–352.
JIME´NEZ, J. A., R. M. ROS, M. J. CANO, AND J. GUERRA. 2002. Contribution
to the bryophyte flora of Morocco: terricolous and saxicolous bryophytes
of the Jbel Bouhalla. Journal of Bryology 24: 243–250.
LAWTON, E. 1971. Moss flora of the Pacific Northwest. Hattori Botanical
Laboratory, Nichinan, Japan.
LISTON, A. 1997. Biogeographic relationships between the Mediterranean and
North American floras: insights from molecular data. Lagascalia 19:
323–330.
LISTON, A., L. H. RIESEBERG, AND T. ELIAS. 1989a. Morphological stasis
and molecular divergence in the intercontinental disjunct genus Datisca
(Datiscaceae). Aliso 12: 525–542.
LISTON, A., L. H. RIESEBERG, AND T. ELIAS. 1989b. Genetic similarity is
high between intercontinental disjunct species of Senecio (Asteraceae).
American Journal of Botany 76: 383–388.
LISTON, A., L. H. RIESEBERG, AND M. A. HANSEN. 1992. Geographic par-
titioning of chloroplast DNA variation in the genus Datisca (Datisca-
ceae). Plant Systematics and Evolution 181: 121–132.
MADDISON, W. P., AND D. R. MADDISON. 1992. MacClade version 3. Analysis
of phylogeny and character evolution. Sinauer, Sunderland, Massachu-
setts, USA.
NYHOLM, E. 1986. Illustrated flora of Nordic mosses. Nordic Bryological
Society, Copenhagen, Denmark.

Page 11

550[Vol. 90AMERICAN JOURNAL OF BOTANY
RAVEN, P. H. 1973. The evolution of Mediterranean floras. In F. Di Cristi and
M. A. Mooney [eds.], Mediterranean type ecosystems: origin and struc-
ture, 213–224. Springer-Verlag, New York, New York, USA.
RAVEN, P. H., AND D. I. AXELROD. 1978. Origin and relationships of the
California flora. University of California Press, Berkeley, California,
USA.
SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER. 2000. ARLEQUIN ver.
2.000. http://anthro.unige.ch/arlequin.
SCHOFIELD, W. B. 1984. Bryogeography of the Pacific coast of North Amer-
ica. Journal of the Hattori Botanical Laboratory 55: 35–43.
SCHOFIELD, W. B. 1988. Bryophyte disjunctions in the Northern Hemisphere:
Europe and North America. Botanical Journal of the Linnaean Society
98: 211–224.
SCHOFIELD, W. B., AND H. A. CRUM. 1972. Disjunctions in bryophytes. An-
nals of the Missouri Botanical Garden 59: 174–202.
SHARP, A. J., H. A. CRUM, AND P. M. ECKEL. 1994. The moss flora of
Mexico. New York Botanical Garden, New York, New York, USA.
SHAW, A. J. 2000. Phylogeny of the Sphagnopsida based on chloroplast and
nuclear and DNA sequences. Bryologist 103: 277–306.
SHAW, A. J. 2001. Biogeographic patterns and cryptic speciation in bryo-
phytes. Journal of Biogeography 28: 253–261.
SHAW, A. J., AND B. H. ALLEN. 2000. Phylogenetic relationships, morpho-
logical incongruence, and geographic speciation in the Fontinalaceae
(Bryophyta). Molecular Systematics and Evolution 16: 225–237.
SMITH, A. J. E. 1978. Moss flora of Britain and Ireland. Cambridge University
Press, Cambridge, UK.
SWOFFORD, D. L. 2001. PAUP*. Phylogenetic analysis using parsimony (*and
other methods). Version 4. Sinauer, Sunderland, Massachusetts, USA.
TAJIMA, F. 1983. Evolutionary relationship of DNA sequences in finite pop-
ulations. Genetics 105: 437–460.
WATTERSON, G. 1975. On the number of segregating sites in genetical
models without recombination. Theoretical Population Biology 7: 256–
276.
WEN, J. 1999. Evolution of eastern Asian and eastern North American dis-
junct distributions in flowering plants. Annual Review of Ecology and
Systematics 30: 421–455.
WOLFE, J. A. 1975. Some aspects of plant geography of the northern during
the late Cretaceous and Tertiary. Annals of the Missouri Botanical Gar-
den 62: 264–279.