12S 16S 12S + 16S Pairwise distances between lineages [%]
T.m. to T.c.c.:
‘Cadiz’ to T.c.c. 6.0 (5.5–7.0) 3.3 (2.6–3.7) 4.9 (4.0–5.5)
‘S.Iberia’ to T.c.c. 5.6 (4.1–7.2) 2.8 (2.1–3.3) 4.6 (3.1–5.2)
‘Portugal’ to T.c.c. 5.8 (5.3–6.8) 3.2 (2.8–3.5) 4.7 (4.0–5.3)
T.m.m. to T.c.c. 5.2 (4.3–6.1) 3.2 (2.1–4.0) 4.4 (3.2–5.0)
‘Gitanilla’ to T.c.c. 5.7 (5.2–6.1) 4.3 (3.7–4.7) 5.0 (4.5–5.4)
T.m.s. to T.c.c. 5.1 (4.6–5.5) 3.0 (2.1–3.5) 4.2 (3.5–4.6)
‘Cadiz’ to ‘S.Iberia’
‘Cadiz’ to ‘Portugal’
‘Cadiz’ to T.m.m.
‘Cadiz’ to ‘Gitanilla’
‘Cadiz’ to T.m.s.
‘S.Iberia’ to ‘Portugal’
‘S.Iberia’ to T.m.m.
‘S.Iberia’ to ‘Gitanilla’
‘S.Iberia’ to T.m.s.
‘Portugal’ to T.m.m.
‘Portugal’ to ‘Gitanilla’
‘Portugal’ to T.m.s.
T.m.m. to ‘Gitanilla’
T.m.m. to T.m.s.
‘Gitanilla’ to T.m.s.
Maximum divergence within each lineage [%]
groups: one represented by T. m. mauritanicus, the other including T. m. simplex as a sister group to samples from
‘Gitanilla’ lineage and south-west Portuguese samples, the
latter two being in a sister-group relationship to each other.
The genetic divergences between ingroup lineages in the
12S and 16S genes and in the dataset with 12S and 16S sequences combined are presented in Table 3. The ‘Cádiz’ and the ‘S.Iberian’ lineage have the lowest inter-lineage sequence divergences in the 16S gene, while their levels of divergence are among the highest observed within Triops mauritanicus in the 12S gene. In contrast, sequence divergences are lowest between ‘Portugal’ and ‘Gitanilla’ lineages in the 12S gene, while these lineages were highly divergent in the 16 gene. Combining the 12S and 16S sequences in a single dataset buffered such effects, which
resulted in an improved estimate of total inter-lineage divergences (assuming that differing branch lengths among these closely related taxa are likely to represent an artefact of short sequences, rather than differences in evolutionary rates). In this combined dataset, uncorrected average inter- lineage p distances between T. c. cancriformis and the lineages within T. mauritanicus range between 4.2 and
5.0%. Uncorrected average inter-lineage p distances ob- served within T. mauritanicus are lower but of similar magnitude, ranging from 2.7 to 3.9%. Intra-lineage diver- gences are highest for T. m. mauritanicus, reaching a maximum of 1.1% (uncorrected p-distance). Thus, within T. mauritanicus uncorrected average inter-lineage p dis- tances were at least 2.5 times higher than the maximum intra-lineage divergence observed, which indicates that all
Fig. 2 Hypotheses on Triops mauritanicus (“T.m.”) and T. c. cancriformis (“T.c.”) phylogeny as reflected by our mitochondrial sequence data; outgroups (Triops longicaudatus, T. granarius, Lepidurus a. apus, L. a. lubbocki, L. arcticus, L. lemmoni) removed for clarity. a ML 12S tree based on large 12S dataset, using TVM+G model of evolution; ML/ MP bootstrap support values given for selected branches. b First of two ML trees based on combined 12S and 16S sequences from selected samples, using GTR+G model; ML
bootstrap support/ Bayesian posterior probabilities given for selected branches. c Strict consensus tree based on combined 12S and 16S sequences (gaps included as fifth character); MP bootstrap support values given for selected branches. All samples labelled with short haplotype names as given in Table A1; for combined datasets,
12S haplotype number given first, followed by blank space and 16S
lineages within T. mauritanicus are well separated from each other.
Geographic distribution of Iberian lineages
In general, suitable habitats for Notostraca form in low- relief landscapes with impermeable surface soils or with
upwelling groundwater (if upwelling lasts for prolonged periods of time after heavy rains). Within the study area, these conditions are predominantly met within the wide valleys of the lower reaches of streams as well as in the coastal lowlands. Thus, the majority of south-west Iberian Triops populations obtained for this study were found at altitudes below 300 meters. However, higher altitudes were
recorded for populations 058 and 059 (Table A1) in
Extremadura (see also Alonso 1985; Pérez-Bote et al.
2006). Distribution records are illustrated in Fig. 3. All four lineages show continuous distribution ranges with sharp range boundaries, and we found no evidence of a mosaic contact zone where different lineages meet.
The ‘S.Iberian’ lineage has the most extensive range, covering wide areas within the valleys of the Guadalquivir and Guadiana rivers and adjacent areas. In south central Portugal it is restricted to the area north of a complex of mountain ranges including the Serra de Monchique and Serra do Caldeirão. The northernmost records are from Cáceres province, Extremadura. The ‘Gitanilla’ lineage appears confined to a small area within the distribution range of the ‘S.Iberian’ lineage and could thus be regarded as sympatric with the latter. However, the minimum geographic distance recorded between the two lineages is
40.5 km. The distribution of ‘Cádiz’ and ‘S.Iberian’ lineages appears to be a typical parapatric one due to the comparatively low minimum distance of 25.8 km recorded between these neighbouring lineages, and to the apparent absence of a geographic barrier between their distribution ranges (if waterbirds are considered as major dispersal vectors, see, e.g., Green and Figuerola 2005). The ‘Portu- gal’ lineage may be geographically more isolated from the
‘S.Iberian’ lineage due to mountain ranges (Serra de Monchique and Serra do Caldeirão) forming its northern range limit, and to a possible lack of suitable habitats towards its eastern range limit, where the mountains approach the sea so that the coastal lowlands are confined
Dispersal abilities inferred from haplotype distribution
For each of 50 populations, we obtained 12S sequences from a minimum of six individuals. Among these popula- tions, at least 52% show associations of multiple haplotypes (percentage of populations with multiple haplotypes was
44% if all populations were considered for which we obtained sequences from at least two individuals each). We identified 18 ingroup haplotypes that occurred in more than a single habitat and thus can be used to infer recent successful dispersal events, i.e. where haplotypes success- fully established in new habitats. Most of these geograph- ically spread haplotypes belong to the ‘S.Iberian’ lineage, for which also the highest accumulative minimum dispersal distances (AMDDs) were observed within Triops maurita- nicus (Table 4), exceeding 200 km in three of the haplotypes. The common haplotype ‘S.Iberia 1’ even dispersed by an accumulative distance of more than
500 km and was detected in a total of 35 different habitats that can only have been reached via overland dispersal. Despite its much smaller range, the ‘Cádiz’ lineage also showed rather high values of AMDD, up to almost 73 km. Inferred successful dispersal was considerably lower for the
‘Portugal’ and ‘Gitanilla’ lineages, with AMDDs of
16.4 km and 1.9 km, respectively. Indeed, for the ‘Portugal’
lineage we found no evidence of recent dispersal among
Fig. 3 Distribution of Triops mauritanicus lineages in south- western Iberian Peninsula, lim- ited to records from this and a preceding study (Korn et al.
2006), as literature records could not be assigned to the phyloge- netic lineages. Black lines show political borders, grey lines the major rivers; dashed area indi- cates extension of marismas (natural temporary marshes) in Guadalquivir River delta around the year 1900
Table 4 Accumulative minimum geographic distances over which single haplotypes must have been passively dispersed to show the
present geographic distribution (marismas of Doñana treated as a single site)
aValues in parentheses indicate total dispersal and total number of sites, respectively, if Triops col- lected from all sampling sites located within the former range of marismas (around year 1900) were treated as separate populations (the temporary lakes and ponds within the marismas are interconnected at certain flood events, so that addi- tional active dispersal may occur among populations within this habitat)