Page 17 - Korn_alii_2006

M. Korn et al. • Sister species within Triops cancriformis

could also be due to low establishment success of dispersed tions, 6.5 Mya, was the trigger for the split of the T. cancriformis
haplotypes: in southern Spain, as well as in Morocco, suitable and T. mauritanicus lineages. The second phase of strong cli-
ecological niches may already have been occupied by locally matic changes coincides with most splits among clades within
adapted populations of native T. mauritanicus subspecies with T. mauritanicus suggesting that these climatic changes might
large resting propagule banks. Thus, new invaders may have have been the trigger for the divergence of these subspecific
to cope with a numerical as well as with a fitness effect, keep- lineages, possibly caused by successive range extensions dur-
ing effective gene flow among populations low (Monopoliza- ing phases with favourable conditions followed by range frag-
tion Hypothesis; De Meester et al. 2002). Although effective mentations and geographical isolation of populations during
long-distance dispersal across the sea may be a scarce event in phases with less favourable conditions (e.g. Veith et al. 2003).
gonochoric populations, there is evidence for its occurrence Calculations with different reconstruction algorithms based
in the present range of Triops species. For example, the distri- on separate 16S and 12S sequence datasets, as well as on the
bution of two closely related species, T. longicaudatus and combined dataset all result in a polytomy between three of
T. australiensis, on different continents, may only be explained the subclades of T. mauritanicus (Fig. 4 and other reconstruc-
by long-distance dispersal. However, for T. cancriformis, we tions, data not shown). This consistent pattern indicates that
could not find any clear indication of gonochoric populations the observed polytomy may not be an artefact caused by
occurring outside of a rather restricted range in northern insufficient data, but rather suggests that a divergence into
Spain (distribution of ‘Spanish T. c. simplex’ in Fig. 2). For this three different clades may have occurred within a very short
species, we hypothesize that a single nongonochoric lineage time. This appears to have been after the first T. mauritanicus
descended from a gonochoric ancestor located in Spain (or lineage (represented by the recent ‘Southern Spanish’ haplo-
possibly Western Europe) that possibly formed parthenoge- type group) had evolved from proto-T. mauritanicus. The fact
netic as well as androdioecious (with a mating system involv- that gonochoric populations of both species as well as the
ing outcrossing and facultative selfing) or hermaphroditic ‘Southern Spanish’ lineage of T. mauritanicus all occur within
populations. This nongonochoric lineage may then have fur- the Iberian Peninsula suggests that the biogeographical ori-
ther diverged into several haplotype groups while spreading gin of both T. cancriformis and T. mauritanicus may be situated
eastwards. Evidence for this hypothesis comes from the first in the Iberian Peninsula. Thus, we hypothesize that a com-
genetic data on reproductive mode (Cesari et al. 2004), the mon ancestor of T. m. mauritanicus, T. m. simplex as well as of
geographical distribution of sex ratios (Scanabissi et al. 2005: the lineages represented by the haplotype groups ‘Portugal’
Fig. 4, Table 2), the distribution of haplotypes in Europe as well and ‘Gitanilla’ was located in the Iberian Peninsula and may
as male occurrence in at least three of the nongonochoric have diverged during a massive range expansion into northern
lineages (see above). This scenario, with only nongonochoric Africa (to both sides of the Atlas Mountains) and other
and thus easily dispersing populations spreading eastwards, parts of Iberia during a phase of especially favourable condi-
might best explain how this rather recently divergent taxon tions. For Triops, such conditions might have been a degrada-
could expand its range to vast areas of Asia, including India tion of formerly dense forests. Actually, a drying trend led to
(Tiwari 1952; Longhurst 1955) and Japan (Suno-Uchi et al. the development and expansion of steppe faunal associations
1997). Most samples described from India (Tiwari 1952), as around 2.3 Mya (Blondel & Aronson 1999), possibly linked to
well as samples from Japan (Suno-Uchi et al. 1997), do not a phase of unusual cold conditions 2.5 Mya (nanoplankton
contain males. Further evidence comes from the apparent data). The strong genetic diversification especially within the
absence of gonochoric populations from mainland Italy lineage represented by ‘Portuguese’ and ‘Gitanilla’ haplotype
(Scanabissi et al. 2005) and Israel (Kuller & Gasith 1996), two groups indicates that the three clades in polytomy evolve
areas that should not have been depopulated during the Ice faster than their sister clade, the ‘Southern Spanish’ haplotype
Ages, and thus could have harboured ancient lineages. group. Molecular data from more populations are needed to
corroborate this possible asynchronous evolution pattern and
Possible triggers for diversification. Triops cancriformis and the biogeographical scenario formulated above.
T. mauritanicus lineages are indicated to have separated 2.3– Within T. cancriformis the different haplotype groups may
8.9 Mya (Table 5). The inferred fluctuations of surface water have diverged during a third period of strong climatic oscil-
temperature in the Mediterranean, based on calcareous lations that occurred since 0.6 Mya and lasts until present
nanoplankton records (fig. 8 in Veith et al. 2003; modified times (as suggested by nanoplankton records, see above). Our
after Müller 1985), suggests that after a long time of constant genetic data suggest that they have diverged around 1.08–
climatic conditions, two periods of strong climatic oscilla- 0.26 Mya, indicating that they were most probably in existence
tions may have occurred in the range of this period. The first well before the last Ice Age. They might thus have formed
occurred during 6.5–5.1 Mya followed by a second one during during repeated range extensions and fragmentations caused
3.9–1.5 Mya. We hypothesize that the onset of these fluctua- by one or several of the preceding glacial cycles.

© 2006 The Authors. Journal compilation © 2006 The Norwegian Academy of Science and Letters • Zoologica Scripta, 35, 4, July 2006, pp301–322 317